University of Ghana http://ugspace.ug.edu.gh UNIVERSITY OF GHANA COLLEGE OF HUMANITIES DEPARTMENT OF ECONOMICS EMPIRICAL INVESTIGATION INTO THE DETERMINANTS OF NON-TRADITIONAL EXPORTS GROWTH IN GHANA: A GRAVITY MODEL OF TRADE APPROACH BY PHILIP APALATOYA (10599198) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE AWARD OF MASTER OF PHILOSOPHY (M. PHIL) DEGREE IN ECONOMICS. JULY, 2018 University of Ghana http://ugspace.ug.edu.gh DECLARATION I, PHILIP APALATOYA, hereby declare that this thesis presented is the original research completed by me under the direction of my supervisors. Neither the entire nor a section has been presented for another degree somewhere else. ……………………………………………… PHILIP APALATOYA (10599198) ………………………………………………. DATE …………………………………. ………………………………. DR. YAW ASANTE PROF. ERIC OSEI-ASSIBEY (SUPERVISOR) (SUPERVISOR) ………………………………….. …………………………………. DATE DATE i University of Ghana http://ugspace.ug.edu.gh ABSTRACT The exports sector plays a vital role in economic growth and development in Ghana. Non- Traditional Exports (NTEs) is essential in expanding the exports sector. Hence, this paper applied the augmented gravity model of trade to investigate the determinants of Non-Traditional Exports (NTEs) growth in Ghana. A panel dataset of Ghana and her 78 major trading partners from 2004 to 2016 was used. The Poisson Pseudo Maximum Likelihood regression (PPML) was used to reveal the effect that variables such as economic size, transportation cost, regional trading blocs, and institutional quality have on Non-Traditional Exports in Ghana. The estimates uncovered that NTEs flows increased significantly with the expansion of variables such as GDP of Ghana, trading partner’s GDP and trading partner’s population. Also, NTEs growth is positively associated with low transportation cost (distance), sharing a common border, common official language and high level of importing country trade openness index. Moreover, high level of the trading partner institutional quality variables such as political stability and absence of violence and rule of law facilitates NTEs positively whereas a high level of regulatory quality control affects NTEs negatively. Regional trading bloc’s variables such as ASEAN, EU, and ECOWAS have significant trade creating potential. Finally, the findings revealed that Ghana’s NTEs have unexploited exports potential with 45 out of the 78 trading partners used in the study. The study recommended that policymakers should implement initiatives that will enhance Ghana’s NTEs to countries with unexhausted trade potentials. With regards to the exhausted trade potential destination, policies that make exports diversification a focal point should be implemented to enable the recapturing of the exhausted markets destination. ii University of Ghana http://ugspace.ug.edu.gh DEDICATION This thesis is dedicated to the Apalatoya and Atampugre families in the consolidation of our family ties. iii University of Ghana http://ugspace.ug.edu.gh ACKNOWLEDGEMENT I thank the Almighty God for his mercies and abundant gift of health and knowledge that have seen me through the completion of this thesis. My special gratitude goes to my academic supervisors Dr Yaw Asante and Dr E. Osei-Assibey who tirelessly guided me throughout this study. Without their effort, I couldn’t have done much by self. My utmost gratitude goes to Madam Victoria Apalatoya (Mother), Mr Jimmy Apalatoya (Father), Madam Stella Atampugre (Aunt), Mr Matthew Atampugre (Uncle) and all my siblings. Also, my gratitude goes to Madam Gina Asumda and all members of the Asumda family for their heartfelt support. Finally, I will like to register my sincere profound gratitude to the Department of Economics at the University of Ghana, my classmates and friends for their moral support and encouragement. iv University of Ghana http://ugspace.ug.edu.gh TABLE OF CONTENT DECLARATION............................................................................................................................ i ABSTRACT ................................................................................................................................... ii DEDICATION.............................................................................................................................. iii ACKNOWLEDGEMENT ........................................................................................................... iv List of Abbreviations ................................................................................................................. viii List of Figures ............................................................................................................................... ix List of Tables ................................................................................................................................. x List of Appendix Tables............................................................................................................... xi CHAPTER ONE ........................................................................................................................... 1 1.1 Introduction .................................................................................................................... 1 1.2 Problem Statement ......................................................................................................... 4 1.3 Objective of the study .................................................................................................... 6 1.4. Research Hypothesis ...................................................................................................... 6 1.5. The significance of the Study......................................................................................... 6 1.6. The scope of the study .................................................................................................... 7 CHAPTER TWO .......................................................................................................................... 8 OVERVIEW OF NON-TRADITIONAL EXPORTS IN GHANA ...................................... 8 2.1. Introduction .................................................................................................................... 8 2.2. Exports in Ghana ........................................................................................................... 8 2.3. Non-Traditional Exports and its Sub-sector ................................................................ 9 2.3.1 Non-Traditional Agricultural Products Sub-Sector .................................................. 9 2.3.2: Processed and Semi-Processed Products Sub-Sector ............................................. 11 2.3.3: Handicraft Sub-Sector ............................................................................................. 12 2.3.4: Trends in Non-Traditional Exports Performance from 1986-2016 ........................ 13 2.3.5 Trend in Non-Traditional Exports Growth Rate from 1987-2016 .......................... 13 2.4. Non-Traditional Exports Markets .............................................................................. 14 v University of Ghana http://ugspace.ug.edu.gh 2.5. Policy Initiatives In Support of Exports growth in Ghana ...................................... 18 CHAPTER THREE .................................................................................................................... 24 LITERATURE REVIEW ....................................................................................................... 24 3.0 Introduction .................................................................................................................. 24 3.1. Theoretical Review on Trade ...................................................................................... 24 3.1.1 Absolute Advantage Theory ................................................................................... 25 3.1.2 Comparative Advantage Theory ............................................................................. 25 3.1.3 Heckscher-Ohlin (H-O) Trade Model ..................................................................... 26 3.1.4 The Specific Factors Model .................................................................................... 27 3.2 Gravity Model of Trade ............................................................................................... 28 3.2.1. Theoretical Justification for the Gravity Equation .................................................. 29 3.3 Empirical Review ......................................................................................................... 32 3.3.1 The Gravity Model of Trade and Exports Performance ......................................... 33 3.2.2 Other models and exports performance ................................................................. 37 3.3. Conclusion .................................................................................................................... 39 CHAPTER FOUR ....................................................................................................................... 40 METHODOLOGY .................................................................................................................. 40 4.0 Introduction .................................................................................................................. 40 4.1. Model Specification ...................................................................................................... 40 4.2. Description of Variables and Data Sourced ............................................................... 42 4.2. 1 Dependent variable ................................................................................................. 42 4.2.2 Independent variables ............................................................................................. 42 4.3. Data Description ........................................................................................................... 46 4.4 Estimation Techniques ................................................................................................. 47 4.4.1 Log Linearization model ......................................................................................... 47 4.4.2 Non-Linear Models ................................................................................................. 49 4.5 Robustness check .......................................................................................................... 53 4.5.1 The Park type test.................................................................................................... 53 4.5.2 Gauss-Newton regression test ................................................................................. 53 4.5.3 Regression Equation Specification Error Test (RESET) ........................................ 54 vi University of Ghana http://ugspace.ug.edu.gh 4.5 Predicting Unexhausted Destination and the Speed of Convergence for Non- Traditional Exports (NTEs) ................................................................................................... 55 CHAPTER FIVE ........................................................................................................................ 57 ESTIMATION AND DISCUSSION OF RESULTS ................................................................ 57 5.0 Introduction .................................................................................................................. 57 5.1 Descriptive Statistics and Correlation Matrix ........................................................... 57 5.2 Diagnostic and Robustness checks .............................................................................. 58 5.2.1 The Time Effect Test .............................................................................................. 58 5.2.2 Heterogeneity Bias .................................................................................................. 59 5.2.3 Heteroscedasticity ................................................................................................... 59 5.2.4: Park Test ................................................................................................................. 60 5.2.5: Gauss-Newton Regression (GNR) tests .................................................................. 60 5.2.7: Regression Equation Specification Error Test (RESET) ........................................ 61 5.3: Discussion of the Empirical Results ............................................................................ 62 5.3.1 Economic Size ........................................................................................................ 65 5.3.2 Distance................................................................................................................... 65 5.3.3 Control Variables .................................................................................................... 66 5.3.4 Regional Trading Bloc Variables............................................................................ 67 5.3.4 Institutional Quality ................................................................................................ 68 5.3 Prediction of Ghana’s Non-Traditional Exports Potential ...................................... 69 5.4 Conclusion ..................................................................................................................... 71 CHAPTER SIX ........................................................................................................................... 73 CONCLUSION AND RECOMMENDATION ........................................................................ 73 6.1 Introduction .................................................................................................................. 73 6. 2 Summary of Findings ................................................................................................... 73 6.3 Recommendations ........................................................................................................ 75 6.4 Limitation of Study ...................................................................................................... 77 References ......................................................................................... Error! Bookmark not defined. Appendix ...................................................................................................................................... 88 vii University of Ghana http://ugspace.ug.edu.gh List of Abbreviations ISSER - Institute of Statistical, Social and Economic Research GEPC - Ghana Exports Promotion Council SAP - Structural Adjustment Policy EPR - Economic Recovery Program NTEs - Non-Traditional Exports PNDC - Provisional National Defence Council viii University of Ghana http://ugspace.ug.edu.gh List of Figures FIGURE 1. 1: TRENDS IN MERCHANDISE TRADE BALANCE OF GHANA FROM 1980-2016 .................. 4 FIGURE 2. 1: TOP TEN PRODUCTS UNDER NTAPS FOR THE YEAR, 2015 ........................................ 10 FIGURE 2. 2: TOP TEN PRODUCTS UNDER PROCESSED AND SEMI-PROCESSED PRODUCTS SUB- SECTOR FOR 2015. .................................................................................................................. 11 FIGURE 2. 3: TOP TEN PRODUCTS UNDER HANDICRAFT SUB-SECTOR FOR 2015 ............................ 12 FIGURE 2. 4: NON-TRADITIONAL EXPORTS PERFORMANCE FROM 1986-2016 ............................... 13 FIGURE 2. 5: NTES GROWTH RATES FROM 1987-2016 .................................................................. 14 FIGURE 2. 6: DESTINATION OF NON-TRADITIONAL EXPORTS BY REGIONAL TRADING BLOCS IN 2016 ............................................................................................................................................... 16 FIGURE 5. 1: OVERALL TOP 10 UNEXHAUSTED EXPORTS DESTINATION FOR GHANA NTES .......... 70 FIGURE 5. 2: OVERALL TOP 10 EXHAUSTED EXPORTS DESTINATION FOR GHANA NTES .............. 70 ix University of Ghana http://ugspace.ug.edu.gh List of Tables TABLE 2- 1: LEADING FIVE COUNTRIES IN EACH MARKET CATEGORY WITH THEIR SHARES OF NTES MARKET WITHIN THE BLOC IN 2016 ........................................................................................ 18 TABLE 5- 1: SUMMARY STATISTICS OF THE DEPENDENT AND INDEPENDENT VARIABLES WITHIN THE PERIOD UNDER STUDY. .................................................................................................... 58 TABLE 5- 2: BREUSCH AND PAGAN LAGRANGIAN MULTIPLIER TEST FOR RANDOM EFFECTS .......... 59 TABLE 5- 3: BREUSCH-PAGAN / COOK-WEISBERG TEST FOR HETEROSKEDASTICITY ..................... 60 TABLE 5- 4: PARK TEST .................................................................................................................. 60 TABLE 5- 5: GAUSS-NEWTON REGRESSION (GNR) TESTS .............................................................. 61 TABLE 5- 6: RESET TEST FOR LOG-LINEAR MODEL ..................................................................... 61 TABLE 5- 7: RESET TEST FOR PPML MODEL ................................................................................ 61 TABLE 5- 8: ESTIMATES OF THE GRAVITY MODEL USING OLS, RE, FE, AND PPML FROM 2004 – 2016 ....................................................................................................................................... 64 TABLE 5- 9: OVERALL GHANA'S MEAN ACTUAL AND POTENTIAL NTES FOR THE ENTIRE PERIOD UNDER STUDY (2004 – 2016). ................................................................................................. 69 x University of Ghana http://ugspace.ug.edu.gh List of Appendix Tables APPENDIX 1. 1: TIME EFFECT OLS ................................................................................................ 88 APPENDIX 1. 2: TIME EFFECT TEST ................................................................................................ 89 APPENDIX 1. 3: PARK-TYPE AUXILIARY REGRESSION AND THE PARK TEST .................................... 89 APPENDIX 1. 4: GAUSS-NEWTON REGRESSION (GNR) TESTS ......................................................... 89 APPENDIX 1. 5: HAUSMAN TEST .................................................................................................... 90 APPENDIX 1. 6 :LIST OF COUNTRIES USED IN THE STUDY WITH THEIR OVERALL ACTUAL, POTENTIAL, UNEXHAUSTED AND SPEED OF CONVERGENCE/DIVERGENCE FOR NTES OVER THE PERIOD UNDER STUDY ......................................................................................................................... 91 APPENDIX 1. 7: CORRELATION MATRIX ......................................................................................... 93 APPENDIX 1. 8: LIST OF COUNTRIES WITH THEIR IMPORT VALUE OF GHANAS’ NTES FROM 2004-2016 .................................................................................................................. 93 xi University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE 1.1 Introduction Exports are considered a vital component in the process of national income determination. They are regarded as an important driver to growth and development in every economy. Exports boost economic growth by generating foreign exchange reserves needed for financing the importation of goods such as energy and investment goods (Eita, 2008). According to Rahman (2009), exports play a crucial role in balancing foreign exchange gap of every economy; hence, they increase a nation’s import capacity and reduce dependence on foreign aid to finance such activities. As noted by Stait (2005), exports promotion can expand intra-industry trade and serve as a support for economies to integrate into the activities of the world economy, hence reducing the impact of external shocks on the domestic economy. The experience of the Latin American and Asian economies highlights how the exports sector is significant for economic growth and development. Given the importance of the exports sector for economic growth in every nation, developing and expanding the exports sector of Ghana’s economy can help to reduce unemployment problems, enhance the balance of payments deficits, increase foreign exchange earnings, decrease heavy external borrowing and subsequently reduce the high level of the poverty rate in Ghana. This thesis intends to contribute to exports expansion in Ghana by investigating factors that determine the growth of Non-Traditional Exports (NTEs). The Ghanaian economy, after independence, was heavily controlled by the state. The country pursued an import substitution strategy aimed at creating a manufacturing base that would encourage production of goods locally and discourage imports. With an effort to restrict imports, the government before 1983, imposed exorbitant import tariffs; three of such tariff schedules 1 University of Ghana http://ugspace.ug.edu.gh were high at 35 %, 60%, and 100% (Nomfundo & Nicholas , 2017). The import substitution strategy was also coupled with a highly protective fixed exchange rate regime. In addition, the overvaluation of the Ghanaian currency at that time rather made imports of intermediate and final goods cheaper, hence having negative repercussion on other sectors of the economy (Oduro et al., 1992). With this economic mismanagement and external market constraints, the economy experienced a continued deterioration in the dollar value of her exports. This reduction in foreign exchange earnings brought about frequent negative growth rate between 1960-1984 with the worst growth rate of negative 14% in 1975 (Fosu, 2001). Hence, it pointed out the risks of depending on the traditional exports base for foreign exchange earnings and stressed the need for the expansion of the exports sector. With efforts to arrest and restore over a decade of economic retrogression, the economy shifted from import-oriented strategies (Import substitution) towards outward orientation (export expansion). As pointed by Medina-Smith (2008), economies that depend on exports expansion development strategies have performed well over the median and long-term than the import substitution strategies. Since the performance of the major exports sector of the Ghanaian economy was unsatisfactory, the monopoly of the major exports was questioned and much attention was given to exports diversification into Non-Traditional Exports (NTEs). The failure of the traditional exports to respond positively to domestic issues and world market price shocks highlighted the need to diversify and broaden Ghana’s exports base to improve upon the growth of the economy. Exports diversification is, therefore, the expansion of the exports base through the addition of new products or markets. It is seen as a critical tool for developing economies to achieve and sustain economic growth and development (Lwesya, 2 University of Ghana http://ugspace.ug.edu.gh 2016). According to Thelle et al. (2015), expanding the market base of an economy through trade enables individual producers to benefit from economies of scale through efficiency in production and the reduction of unit cost of production necessary for sustained economic growth. With efforts to diversify exports, a policy such as the Economic Recovery Program (ERP) was implemented in 1983. The focal aim of this program was to boost the domestic market through exports promotion to halt the negative growth and stabilize the economy on a reasonable track (Fosu, 2001). As noted by ISSER (2006), the aim was to make exports promotion and diversification a focal point. In 1986, the Structural Adjustment Policy (SAP) was also implemented and it was to serve as a supplement to the ERP. The NTEs sector was accorded an unparalleled attention in attaining economic growth and development under the ERP and SAP. Non-Traditional Exports (NTEs) refer to exports outside the principal or traditional export products (Osei-Assibey, 2015). The sector is an important contributor to Ghana’s overall exports in recent years and is driven by three main sub-sectors: agricultural, manufacturing and the handicraft sector. The NTEs are exported to 137 countries. These countries have been divided into five groups such as the European Union, ECOWAS, other developed countries, other African counties and other with markets shares of 35.84%, 31.59%, 3.62%, 7.52% and 21.43% respectively (GEPC, 2015). The expansion of NTEs have numerous advantages such as improvement of employment level, increase in foreign exchange earnings, introduction of new technology and capacity building, diversifying the economy both at the village and national levels, and the stimulation of private entrepreneurship in the economy (Ampadu-Agyei, 1994). As noted by Thrupp et al. (1995), NTEs contributed to rapid growth in export earnings. The contribution of NTEs to total exports 3 University of Ghana http://ugspace.ug.edu.gh of Ghana in 1987 was valued at US$ 27.96 million and then increased to US$ 2,462.85 million in 2016. This performance of NTEs sector might seem very good; however, the growth rates of NTEs have been fluctuating over the years. The growth rates of NTEs from 2010 to 2016 were 34.08%, 48.73%, -2.43%, 3.04%, 3.19%, 0.31% and -3.24% respectively (GEPC, 2016). Trends and growth rates of NTEs in Ghana are presented in Figure 2.4 and 2.5 respectively of Chapter Two respectively. 1.2 Problem Statement The government of Ghana over the years has made efforts to improve upon foreign exchange earnings through the pursuant of tangible trade policy measures, incentive schemes, and diversifying exports into non-traditional products. However, data from the World Bank show that foreign exchange earnings for the economy have not been satisfactory over the years. Figure 1.1 shows that Ghana’s economy has been experiencing negative merchandise trade balance from 1988 to 2016. The merchandise trade deficit in Figure 1.1 is attributed to the inability of the merchandise exports sector in providing enough foreign exchange earnings to finance merchandise imports of the economy. Figure 1.1: Trends in Merchandise trade balance of Ghana from 1980-2016 1000 0 -1000 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 -2000 -3000 -4000 -5000 -6000 Source: Own estimated values using data from World Bank, 2016 4 University of Ghana http://ugspace.ug.edu.gh The outcome on the merchandise exports sector indicates that NTEs sector over the years has not been very successful in fulfilling its main objective of serving as a shield to the vulnerable economy against the instability nature of earnings from the major traditional exports. Although, Figure 2.5 of Chapter Two shows that NTEs have experienced significant growth rates over the years with top four growth rates of 363.27%, 66.40%, 73.01%, and 48.75% recorded in 1987, 1994, 1996 and 2011 respectively, NTEs have witnessed negative growth rates with the worse four negative growth rates of 61.42%, 18.02%, 9.05% and 2.33% recorded in 1988, 1989, 2009, and 2016 respectively. The trends in the growth rates indicate that key issues confronting NTEs growth have not been identified with certainty, hence, more research needs to be undertaken in order to provide tangible policy solutions to NTEs issues. Several studies have been conducted in Ghana by different researchers such as Kyereboah-Coleman & Biekpe, (2006), Oduro et al., (1994), and Addo et al., (1999) aimed at promoting NTEs growth. A deeper look into a more recent literature such as Kyereboah-Coleman & Biekpe (2006) revealed that their work concentrated on the determinants of NTEs at the firm level and this limited their scope of the study to only internal factors but factors that affect NTEs are both internal and external. Furthermore, the inability of the works cited above in identifying certainty key factors that promote NTEs growth in Ghana could be attributed to their methodological choices. None of the studies above applied the gravity model, but, this model is regarded as one of the most successful and widespread methodology used in estimating determinants of international trade flows. Unlike other models, the gravity model allows for more variables to be taken into account when examining the degree and forms of international trade (Paas, 2000). According to Echengreen & 5 University of Ghana http://ugspace.ug.edu.gh Irwin (1998), the gravity model of trade is considered as the workhorse of empirical studies of bilateral trade to the virtual exclusion of other approaches. Given the shortfalls of the studies cited above, this study is being carried out to determine both the external and internal factors that affect the growth of NTEs in Ghana using the gravity model of trade. 1.3 Objective of the study The main objective of this study is to investigate the determinants of NTEs growth in Ghana The specific objectives of this study are: i. To determine the internal and external factors that affect the growth of Non- Traditional Exports (NTEs). ii. To investigate exports destination potential available for Ghana’s NTEs. 1.4. Research Hypothesis i. Internal and external factors do not affect the growth of non-traditional agricultural exports. ii. There is no exports destination potential available for Ghana’s NTEs. 1.5. The significance of the Study The government of Ghana trade policy aim since 1983 is to reduce dependence on traditional exports by diversifying into Non-Traditional Exports. The results of this study are very crucial in notifying the government of the necessary steps to be undertaken to boost the exports growth of 6 University of Ghana http://ugspace.ug.edu.gh Non-Traditional Products. Also, knowledge on determinants of NTEs has a positive ramification on producers and farmers welfare, since the expansion of the sector will increase their income levels. Furthermore, the production of NTEs Products is mostly labour intensive; hence expansion of the sector will increase the employment level. Finally, the study is going to add to the growing literature on the determinants of international trade. 1.6. The scope of the study The study examines determinants of Ghana’s NTEs growth using annual panel data set of Ghana and 78 major trading partners from 2004 to 2016. The selection of the period is due to data availability. Also, because panel analysis was carried out, the research concentrates mainly on those exports that are registered under the Ghana Export Promotion Council (GEPC) as NTEs. Such products have available data needed for the analysis. 1.7. Organization of the study This study is going to be divided into six main chapters. Chapter one is devoted to the introduction of the study and outlines the background, problem statement, objectives, hypotheses, significance and scope of the study. Chapter two presents an overview of the NTEs sector in Ghana. Chapter three is devoted to reviewing of related literature. The fourth chapter focuses on the methodology which presents the empirical model and econometric estimation technique employed in attaining the objectives of the study. Chapter five presents the econometric estimation results and discusses the characteristics of the panel dataset and finally, chapter six presents the summary, conclusions and recommendations. 7 University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO OVERVIEW OF NON-TRADITIONAL EXPORTS IN GHANA 2.1. Introduction The chapter takes a careful look at exports categories in Ghana, an overview of the overall non- traditional exports sector and its subsectors in Ghana. Specifically, the trend analysis of overall non-traditional exports and its subsectors in Ghana since 1986 to 2016, Ghana’s major export markets destination and various policies initiatives were taken to promote exports growth in Ghana shall be provided in this section. 2.2. Exports in Ghana The exports in Ghana are grouped into traditional and non-traditional exports. The traditional exports comprise primary commodities or raw materials, for example, cocoa beans, minerals (gold, diamond, bauxite and manganese) and timber. The Non-Traditional Exports (NTEs) are exports of commodities outside the traditional exports and comprise of commodities such as handicraft, aluminium products and horticulture products (GEPC, 1986). The NTEs sector is regarded as key to poverty reduction, especially in the rural communities in Ghana due to its job creation opportunities and income generation. The traditional exports sector is the major contributor to the exports earning and it contributes about 80% to the exports earnings in the economy. The traditional and non-traditional exports contributed 81% and 19% respectively to the total exports earnings in 2015 (GEPC, 2015). 8 University of Ghana http://ugspace.ug.edu.gh 2.3. Non-Traditional Exports and its Sub-sector As noted earlier, any exports of product outside the principal exports are termed as Non- Traditional Exports (NTEs). Some examples of products that currently make the NTEs basket in Ghana are cashew nuts, medicinal plant and parts, fresh or chilled tuna, pineapples, bananas, mangoes, papaya, citrus, shea nuts, cocoa paste, articles of plastic, canned tuna, machinery and parts, veneer, hides and skins, ceramic products, traditional musical instruments, kente products and beads (ISSER, 2013). The Non-Traditional Exports (NTEs) were initiated in 1969 after the setting up of the Ghana Export Promotion Authority (GEPA). Its introduction was a deliberate policy of Government aimed to diversify the traditional export sector to increase exports contribution to GDP growth, cushion the economy of Ghana against the instability in prices of traditional commodities exports and help to solve the frequent negative trade balance. The NTEs sector is driven by three subsectors namely agricultural, processed and semi-processed and handicrafts. 2.3.1 Non-Traditional Agricultural Products Sub-Sector The groups of products that are categorized under non-traditional agricultural subsector are horticulture, fish and seafood, game and wildlife, cereals, oilseed and nuts, and other agricultural products. Examples of products under this subsector are pineapples, mangoes, cashew nuts and oil seed, bananas, medicinal plant and parts, tropical flowers and vegetables, yam, natural rubber, kola nuts, cotton seeds, maize, coconuts, lobsters/shrimps/prawns, textile and garments shea nuts and so on. The top ten products that drive the performance of the agricultural sub-sector in 2015 are medicinal plants and parts, fresh or chilled tuna, cashew nuts, shea nuts, banana, pineapples, yams, mangoes, other fresh or chilled fish, palm nuts and kernel. 9 University of Ghana http://ugspace.ug.edu.gh Figure 2.1 below shows the exports value of the top ten products under the Non-Traditional Agricultural Products (NTAPs) for the year 2015: Figure 2. 1: Top Ten Products under NTAPs for the Year, 2015 250 211.34 200 150 100 50 33.57 29 27.78 25.44 20.54 19.98 7.62 3.7 3.34 0 Source: Own estimated values using 2015 data from GEPC Figure 2.1 shows the exports value of the top ten products under the NTAPs, the highest being cashew nuts with exports value of US$ 211.32 million and 56.99% as the growth rate. Palm nuts and kernel was the least contributor among the top ten, its exports value and growth rate for the year were US$ 3.3 million and 88.03% respectively. Earnings of the agricultural exports sub- sector from 2014 to 2016 were US$340.68 million, US$396.96 million, and US$371.14 million, which represents a growth rate of 5.25%, 16.5%, and -6.5% respectively. The sub-sector contributed 15.07% to total NTEs earning in 2016 (GEPC, 2016). 10 VALUE (M. $) University of Ghana http://ugspace.ug.edu.gh 2.3.2: Processed and Semi-Processed Products Sub-Sector Products that make the basket of Processed and semi-processed subsector are wood products, manufacturing products such as canned food and beverages, electrical cables and aluminium, pharmaceuticals and other processed products. The top ten products that influenced the performance of this sub-sector in 2015 were cocoa paste, canned tuna, articles of plastic, cocoa powder, lubricating oil, shea oil, aluminium plates and sheets, machine parts, natural rubber sheets and freshly cut fruits. Figure 2.2 below shows the exports value of the top ten products under processed and semi-processed subsector in 2015. Figure 2. 2: Top ten Products under Processed and Semi-Processed Products Sub-Sector for 2015. 600 510.72 500 400 300 219.01 183.6 200 107.99 82.53 100 64.03 61.58 59.22 53.52 48.63 0 Source: Own estimated values using 2015 data from GEPC Figure 2.2 above shows the exports value of the top ten products under the Processed and Semi- Processed Sub-Sector, the highest being cocoa paste with an exported value of US$ 510.72 million and -10.21% as its growth rate whereas the least product being freshly cut fruit with exported value and growth rate for 2015 as US$ 48.63 million and -3.11% respectively. Earnings of the Processed and Semi-Processed Sub-Sector from 2014 to 2016 were US$2,169.66 million, 11 VALUE ( MILLION US$) University of Ghana http://ugspace.ug.edu.gh US$2,120.5 million, and US$2,086.49 million, which represents a growth rate of 2.83%, -2.27%, and -1.60% respectively. The sub-sector contributed 84.7% to total NTEs earning in 2016 (GEPC, 2016). 2.3.3: Handicraft Sub-Sector The handicraft subsector entails mainly of products such as carving and weaving woodcrafts, beads, jewellery and services such as medicinal tourism, financial services and. The top ten products that influenced the performance of the handicraft exports sub-sector in 2015 are batik/tye and dye, basket-ware, hides and skins, traditional musical instruments, other handicraft items, statuettes beads paintings and drawings, and articles of jewellery. Figure 2. 3: Top Ten Products under Handicraft Sub-sector For 2015 1.6 1.37 1.4 1.2 1 0.82 0.8 0.73 0.6 0.4 0.4 0.25 0.25 0.2 0.13 0.1 0.081 0.066 0 Source: Own estimated values using 2015 data from GEPC Figure 2.3 above shows the exports value of the top ten products under the handicraft sub-sector, the highest being batik/tye and dye with an exported value of US$ 1.37 million and 381.58% as its growth rate while the least in this subsector being Article of jewellery with exported value and 12 VALUE( US $MILLION ) University of Ghana http://ugspace.ug.edu.gh growth rate in 2015 as US$ 0.066 million and 582.07% respectively. Earnings of the handicraft exports sub-sector from 2014 to 2016 were US$3.48 million, US$4.27 million, and US$5.22 million, which represent growth rates of 41.46%, 22.70%, and 22.25% respectively. The sub- sector contributed 0.21% to total NTEs earning in 2016 (GEPC, 2016). 2.3.4: Trends in Non-Traditional Exports Performance from 1986-2016 Figure 2.4 shows that Ghana’s NTEs has witnessed a continuous increase in its exports values over the past years, it climbed roughly from USS $ 23.74 million in 1986 to a value of USS$ 2.462.85 billion in 2016. This corresponds to the yearly average growth rate of 25.77 % from 1986-2016. Figure 2. 4: Non-Traditional Exports Performance from 1986-2016 3000.00 2500.00 2000.00 1500.00 1000.00 500.00 0.00 Source: Own estimated values using 1986-2016 data from GEPC 2.3.5 Trend in Non-Traditional Exports Growth Rate from 1987-2016 Figure 2.5 below shows the annual growth rates of Non-Traditional Exports (NTEs) of Ghana from 1987-2016: 13 Value(USS$ Million) 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 University of Ghana http://ugspace.ug.edu.gh Figure 2. 5: NTEs Growth Rates from 1987-2016 400 350 300 250 200 150 100 50 0 -50 -100 Source: Own estimation using data from Figure 2.4. Figure 2.5 above demonstrates that, although, there were achieved growths in some years, NTEs growth rates in Ghana have witnessed continuous fluctuations over the years. In 1987, the NTEs achieved a growth rate of 363.27%, this growth rate could be attributed to the massive economic reforms that took place from 1982-1986. However, this was followed by negative growth rates of 61.42% and 18.02% in 1988 and 1989 respectively. Again, NTEs had a growth rate of 48.75% in 2011 but negative growth rates of -2.43% and -2.33% in 2012 and 2016 respectively. This shows that NTEs performance is not very satisfactory and there is the need to enhance its growth rate in a positive direction. 2.4. Non-Traditional Exports Markets Given the downward trends in exports growth between 1970 and 1982, the adoption of an outward-looking strategy (trade openness) through the implementation of trade liberalization policies in 1983 brought an upward trend in trade values. The value of exports which stood at GHȻ 10,223.2 million in 1984 increased to GHȻ 94,437 million by the end of 1986 and surged to GHȻ 284,266.1 million in 1990 (Oduro et al, 1992). Ghana continued to expand trade by 14 NTES Growth Rate( Percentages) 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 University of Ghana http://ugspace.ug.edu.gh primarily trading with countries such as Britain and Germany. Britain was the highest destination of Ghana’s cocoa beans, absorbing about 50% of the total cocoa beans exported. In 1992, Germany was the top destination country for Ghana’s exports; it absorbed approximately 19% of all exports from Ghana, followed by Britain, also accounting for about 12%, and the United States and Japan accounting for 9% and 5% respectively (GEPC, 2015). EU and ECOWAS are the top marketing destination for Ghana’s NTEs over the years. In 2010, the EU import of NTEs from Ghana amounted to US$863.15 million representing 52.98% of the NTEs Market shares, followed by ECOWAS, Asia, US and other countries with imports of NTEs amounting to US$430.35 million, US$129.59 million, US%119.65 million and US$86.32 million representing market shares of 26.41% ,7.97 ,7.34% and 5.30% respectively. The performance of EU was attributed to Netherlands, France, and UK increased in imports of NTEs shares by 15.35%, 11.21% and 8.01% respectively, whereas that of the ECOWAS was due to Togo, Nigeria and Burkina Faso increase in imports of NTEs shares by of 7.26%, 5.91% and 5.18% respectively (GEPC, 2010). Furthermore, EU share of NTEs from Ghana increased to US$1,070.75 million representing 42.59% and then reduced to US$1039.47 million representing 41.22% in 2014 and 2015 respectively. This reduction in the share of EU was due to a negative growth rate of 16.82% from the Netherlands market share. ECOWAS share of NTEs from Ghana surged to US$759.56 million representing 30.21% and then increase to US$796.60 million representing 30.12% in 2014 and 2015 respectively. The increased in ECOWAS share of the market was due to positive market share growth rates of 82.86% and 31.92% between Togo and Mali respectively. The United States of America market shares of Ghana’s NTEs had a growth rate of -2.38% between 2014 and 2015 (GEPC, 2015). Figure 2.6 below shows that ECOWAS sub-region is the largest destination for Ghana’s NTEs absorbing about 37% of NTEs market 15 University of Ghana http://ugspace.ug.edu.gh shares ,followed by European Union (EU), Asia, North American, other Countries and other African countries in that order with marketing shares of 35%, 18%, 5%, 3% and 2% respectively in 2016 (GEPC, 2016). Figure 2. 6: Destination of Non-Traditional Exports by regional trading blocs in 2016 5% 3% 2% 18% 35% EUROPEAN UNION ECOWAS ASIA OTHER AFRICA COUNTRIES NORTH AMERICAN OTHER COUNTRIES 37% Source: Own estimated values using 2016 data from GEPC. The top ten countries of destination for NTEs from Ghana in 2016 were Burkina Faso, United Kingdom, Vietnam, Netherlands, Togo, Nigeria, France, Mali, United States, and Spain. Figure 2.7 below shows the top ten countries with the highest imports value of NTES from Ghana. From Figure 2.7, the top and least importing country for NTEs in 2016 are Burkina Faso and Spain with imports values of US$248.18 and US$108.22 representing a growth rate of 24.26% and 49.72% respectively. Although France, Netherlands and Togo are part of the top ten countries, their market shares witnessed a reduction in the growth of 27.85%, 37.76%, and 33.73% respectively as compared to values in 2015. 16 University of Ghana http://ugspace.ug.edu.gh Figure 2.7: Top ten countries with highest imports values of NTEs from Ghana in 2016 300.00 248.18 250.00 203.29 200.00 154.65 150.94 146.60 142.30 150.00 124.81 122.40 113.62 108.22 100.00 50.00 0.00 Source: Own estimated values using 2016 data from GEPC Finally, Table 2-1 below shows the top five countries within each market categories with the United Kingdom, Burkina Faso, Vietnam, United States, Switzerland and Egypt being top countries within European Union, ECOWAS, Asia, North American, other countries and other African countries respectively in 2016. 17 (US$ Million) University of Ghana http://ugspace.ug.edu.gh Table 2- 1: Leading Five Countries in each Market Category with their shares of the NTEs market within the Bloc in 2016 MARKET CATEGORY COUNTRY MARKET SHARES (%) EUROPEAN UNION UNITED KINGDOM 23.38 NETHERLANDS 17.36 FRANCE 14.35 SPAIN 12.45 BELGIUM 8.02 ECOWAS BURKINA FASO 27.05 TOGO 15.98 NIGERIA 15.51 MALI 13.34 SENEGAL 10.08 ASIA VIETNAM 35.50 INDIA 19.32 MALAYSIA 12.54 CHINA 7.98 SINGAPORE 4.18 NORTH AMERICAN UNITED STATES 87.41 CANADA 9.80 BAHAMAS 1.35 MEXICO 1.02 TRINIDAD AND TOBAGO 0.30 OTHER COUNTRIES SWITZERLAND 63.25 NORWAY 16.32 AUSTRALIA 12.86 BRAZIL 6.10 ICELAND 0.39 OTHER AFRICAN EGYPT 28.79 COUNTRIES SOUTH AFRICA 15.22 MOROCCO 13.20 CAMEROON 10.48 CONGO 9.53 Source: Own estimates using 2016 data from GEPC, 2016 2.5. Policy Initiatives In Support of Exports growth in Ghana After independence, successive governments in Ghana have pursued varying degrees of trade policies, programmes and projects aimed at increasing the growth of the exports sector to enable the sector to increase its competitiveness level in international markets. These policies, 18 University of Ghana http://ugspace.ug.edu.gh programmes and projects were factored in the national developmental framework of the economy. As noted earlier, the need to promote exports growth in the economy made exports diversification focal points since the 1960s. This desire brought about the establishment of Ghana Exports Promotion Council (GEPC) in 1969. The aim of GEPC was to encourage, assist and develop the production of exportable products in a manner which the council deemed necessary. Exports expansion initiatives and import restrictions were introduced to boost NTEs especially the manufacturing subsector. Some of the initiatives that were implemented to encourage the exports of domestic manufacturing products comprised income tax rebates, local tax waivers, exports bonus, and automatic import licenses renewal for raw materials that were important to the manufacturing firms. As pointed out by Oduro et al (1992), the impact of those exports initiatives and policies were limited due to the failure to incorporate all components necessary to the exports promotion packages to ensure a consistent response from the NTEs. Due to these challenges, the Provisional National Defence Council (PNDC) government at that time was committed to make NTEs promotion a key priority within the context of the ERP in 1983. This initiative brought about a significant change in objectives, emphasis and the pattern of national development policy. Given government commitment to enhanced exports expansion in 1983, new policies and programs aimed at removing constraints such as fragile production base, high domestic production cost, non-availability of funding’s, poor marketing infrastructures, insufficient exports incentives, lack of information about exports procedures and weak institutional supports confronting the NTEs sector (Addo & Marshall, 1998). Also, institutions 19 University of Ghana http://ugspace.ug.edu.gh which were supposed to assist the government and other parties involved in NTEs operations to overcome the constraints were also unable to carry this function because they did not have the adequate resources that are necessary to fulfil that responsibility. Furthermore, several operators complained of the insignificance of research institutions since research findings concerning product development usually reached them either too late to be of value or too expensive for many scale smallholders. These institutional and research challenges compelled institutional reforms and made the establishment of export awareness initiatives throughout the country a government priority. These reforms were introduced to ensure that the financial needs of the NTEs were adequately met, incentives were adequately provided, marketing, infrastructure were adequately improved, funding was made readily available and export procedures were streamlined. As part of the export reforms, the Ministry of Trade and Tourism (MOTT) and GEPC were restructured to enable export administration and documentation. At the same time, the financial sector also witnessed major reforms aimed at streamlining their transactions both nationally and internationally in supports of NTEs (Addo & Marshall, 1998). As noted by GEPC (1987), the most significant initiatives to the NTEs were the formation of Export Finance Company (EFC) in 1989 which sought to resolve the problem of inadequate export financing. Export incentives were therefore made available to exporters and producers of exportable products to:  Attract investment to the NTEs and boost exports development.  Reduce the cost of production of export products and make them very competitive at international markets. 20 University of Ghana http://ugspace.ug.edu.gh  Enable effective administration of the incentive system st Also, the Free Zones Act (1995), Act 504 was passed on 31 August 1995 by the parliament of Ghana to ensure fast exploitation of the country’s exports potential. The key reason why the Free Zones Board was set up was to monitor and support the activities of the Exports Processing Zones (EPZs) to be established in the country. Its key objectives were to attract foreign direct investments, provide employment opportunities, increasing foreign exchange earnings, provide business opportunities to enable both local and foreign investors to undertake joint-ventures enhance technical know-how of Ghanaian and diversify exports. In 1998, Government of Ghana launched the gateway programme to promote Ghana as the trade and investment centre of West Africa. The main areas of focus for the Ghana Free Zone Board were enclave in the development of manufacturing, service, and warehousing for NTEs products (Darko-Opoku, 2014). The government of Ghana implemented the National Exports Strategy (NES) in 2012. The key objective of this trade initiative was to put Ghana as a world-class exporter of competitive products and achieve a significant increase in the NTEs with annual growth targeting of about th US$5.00 billion in its 5 year in 2017. In this regard, officials of the Ministry of Trade and Industry (MoTI), Ministry of Finance (MoF), Ghana Exports Promotion Authority (GEPA) and Other Private Associations under the national Exports development to formed working group that would identify and implement development projects that would enable the expansion of NTEs sector (Ghana Ministry of Trade and Industry, 2013). 21 University of Ghana http://ugspace.ug.edu.gh The African Growth and Opportunity Act (AGOA) is another export enhancing program that Ghana has benefited from as a member. It is a U.S. preferential trade program targeted to help spur export-led growth and development in Sub-Saharan Africa and also, enhance U.S. trade and investment ties within the region. The African Growth and Opportunity Act (AGOA) was passed in 2000 by the U.S. Congress. Ghana together with other 43 Sub-Saharan African countries are currently members of AGOA. The economy has benefited from this trade program after being declared as a member in 2000. AGOA was initially due to expire in 2008, but, later extended to 2015 and it is now set to expire in 2025. As noted in the previous chapter, the aim of the Government of Ghana is to pursue a development strategy based on export growth and expand inward direct investment in the economy. This led to the development of the National Trade Policy in 2005. This trade policy aimed to increase the country’s international competitiveness level and ensure greater market access for products from Ghana. In sum, the initiative seeks to promote regional trade integration. Ghana benefited enormously in the production and exports of both traditional and NTEs from these trade initiatives, but as noted by the Ministry of Trade, Ghana has not fully utilized this opportunity to expand exports to the U.S as compared to other African countries. In 2014, Ghana exports to the U.S. amounted to US$ 1,162,825 thousand but it reduced to US$ 914,962 thousand and US$808,307 thousand between 2015 and 2016 respectively (U.S. Department of Commerce, 2016). According to Federation of Association of Ghanaian Exporters (FAGE), Ghana has not 22 University of Ghana http://ugspace.ug.edu.gh made the most of this AGOA opportunity due to its inability to build the needed production capacity to meet the demand requirements from AGOA and other trade organizations. Another trade initiative that presented the Government of Ghana the opportunity to boost the growth potential of NTEs is the Economic Partnership Agreement (EPA). This is an agreement between ECOWAS and EU countries. The negotiations of EPA between ECOWAS and the European Union was started in October 2003 but, could not be concluded in December 2007 as specified in the Cotonou Agreement (cf. Article 37.1), nor even at the end of October 2009 as agreed on 17th June 2007 by the Chief Negotiators of the two regions. Negotiations of EPA between the two contracting parties were finally concluded in Brussels in 2014. The implementation of the EPA initiative was critical to ensure that Ghana’s exports to the EU were not disrupted and to enable the protection of the agricultural sector against devastation and unfair trade practices (Opoku, 2015). 23 University of Ghana http://ugspace.ug.edu.gh CHAPTER THREE LITERATURE REVIEW 3.0 Introduction This chapter is structured into three sub-sections: Theoretical Review, Methodological Review (Gravity model of trade), and the Empirical Review. 3.1. Theoretical Review on Trade Trade is considered a critical tool in economic growth and development in every economy. Thus, it serves as a vital instrument for economic development, since; it provides foreign exchange essential for industrialization in every economy (Rahman, 2009). Modern theories of trade are due to sequential development of ideas in economic thought. Thus, trade theories are attributed to the result of the advancement of economic thought formed from the writings of philosophers from the mercantilists and later that of the classical such as Adam Smith, David Ricardo and John Stuart Mill. Their writings played a significant part in shaping modern trade theories. The rationale behind modern trade theories is to scrutinize the causes, the types of trade that economies are involved, and the benefits of trade on the welfare of people in these economies. Since, the early views on trade by classical theorists form the foundation for contemporary trade theory, and some of their views still influence modern trade theory, the study will present a brief overview on some of these classical trade theories such as absolute advantage model, theory of comparative advantage, the Heckscher-Ohlin model and the specific factors model in terms of their views on the causes and importance of international trade. 24 University of Ghana http://ugspace.ug.edu.gh 3.1.1 Absolute Advantage Theory Adam Smith (1723-1790) propounded the theory of absolute advantage trade model in his book titled, the Wealth of Nations (Smith, 1776). It was a reaction against the mercantilist’s static views on the cause and benefit of trade in the world economy. Smith showed that the possible benefit from specialization is not associated with the division of assignment among workers inside firms only, but, trade between countries as well. His absolute advantage theory was against the strong restriction of foreign-trade advocated by the mercantilists and rather advocated for free trade. Smith noted that the rationale behind trade between nations is due to the absolute advantage in the production of commodities, and it exists, when a nation has a superior advantage in the production of a particular good than others. From this viewpoint, the theory postulated that a nation should export products in which it is more superior in production and import the ones that it is less superior in production. Smith argued that with free trade, world output will increase and both countries benefit after trading with each other (Carbaugh, 2012; Dunn & Mutti, 2004 and Salvatore, 1998). However, Smith’s theory was not able to clarify why nations which have a complete advantage in the production of all the tradeable products still trade with different nations which have an absolute disadvantage in the production of such products (Carbaugh, 2012; Dunn and Mutti, 2004 and Salvatore, 1998). 3.1.2 Comparative Advantage Theory Due to the looseness in the absolute advantage trade theory concerning how gains from trade are possible, David Ricardo (1772-1823) propounded the theory of comparative advantage as an extension to the absolute advantage theory, and, showed that mutually beneficial to trade is still possible whether or not absolute advantage exists between nations. The theory states that a 25 University of Ghana http://ugspace.ug.edu.gh nation should specialize and export goods in which it has a lower opportunity cost in its production whereas, the other nation should specialize and export goods which it has a lower opportunity cost in production. Hence, the world output of the goods would increase. As noted by Anderson (2004) and Suranovic (2006), the comparative advantage theory maintains that international difference in productivity of workers is the main reasons behind international trade. The empirical work of MacDougall (1951), Stern (1962), and Balassa (1963) affirm the Ricardian conclusion that unequal distribution of the factors of production (labour) is the basis for comparative advantage. However, this model came under serious criticism for its unrealistic assumptions and inability to explain clearly neither the reasons behind the unequal distribution of labour productivity across countries nor the impacts of trade on the welfare of the factors of production (Salvatore, 1998). 3.1.3 Heckscher-Ohlin (H-O) Trade Model Heckscher-Ohlin (H-O) trade model was developed by two economists from Sweden namely Eli Heckscher (1919) and Bertil Ohlin (1933) as a modification to the comparative advantage model by introducing capital as an additional input into production. The rationale behind the extension of the comparative advantage theory was to examine the reason behind international differences in labour productivity, factors that define the degrees of comparative advantage between countries and the forms of international trade flows. The authors argued that comparative advantage arises due to unequal distribution of productive resource or factor endowments between nations. Hence, from their point of view, the more abundant a factor is to an economy, the lesser the cost associated with its usage, hence, providing the economy with the advantage to adopting production technique that will intensively make use of the more abundant factor. They 26 University of Ghana http://ugspace.ug.edu.gh assumed that distinctive products demand different levels of factor input usage in their production, hence, the model postulated that nations should export products that make intensive use of those factors that are domestically abundant, and import products that make intensive usage of those factors that are domestically un-abundant (Hill, 2009 and Salvatore, 1998). 3.1.4 The Specific Factors Model The specific factors model varies very slightly from the Heckscher-Ohlin trade model. The main difference is that the specific factors model is considered as a short run whiles the Heckscher- Ohlin model is regarded as a long-run model. The model assumes that every nation is endowed with three distinct factors of production namely labour, capital and land. Furthermore, the model assumes that, the economy produces manufacturing and agricultural goods, and that labour is perfectly mobile across sectors but capital and land are specific to the manufacturing and agricultural industries respectively. Given these suppositions, the model arrived at a key conclusion that, the pattern of international trade flows is influenced by the relative abundance of the specific factor. With regards to the benefits of free trade, the model concludes that the benefit to the owners of the mobile factor is ambiguous, thus, the consumption preference of the labour determines the impact of trade on the owners of labour. For example, a labour whose consumption preference is biased toward the exportable product may be worse off, whereas, labour whose consumption choice is for the importable product benefits. On the other hand, the effects of free trade on the proprietors of the specific factors are unambiguously determined. Thus, the proprietor of a specific factor used in the production of an exportable product becomes well off after trade while the proprietor of a specific factor of the importing competing sector becomes worse off. 27 University of Ghana http://ugspace.ug.edu.gh 3.2 Gravity Model of Trade Given, the looseness of the complete specialization trade models cited above in clarifying the degree of trade between nations, the gravity theory is considered to be exceptionally effective in this regard. Thus, trade theories cited above only explain why nations trade in diverse products but do not explain why the degree of trade relations differs between nations. However, the gravity model makes it possible for more variables to be added to explain the degree of trade as part of international trade flows (Paas, 2000). In recent years, the gravity model of trade is viewed as the best and broad trade model used in assessing the determinants of international trade flows. The gravity model of trade is a reminiscence of the Newton theory of gravitation (Becchetta et al, 2012). The Newton theory of gravitation postulates that two bodies attract each other in proportion to the product of each body’s mass (in kilogram) divided by the square of the distance between their centres of gravity(in meters). Stewart (1948) and Zipf (1946) utilized the gravity model in the social sciences to study trips among urban area with the assistance of the following specification: (3.1) Where represents the number of journeys between city i and j, POP is the population in city i and j, represents the distance between the cities and E is the intercept of the equation (Rahman, 2003; Chritie, 2002; Zhang & Kristensen, 1995). With respect to trade, Tinbergen (1962) and Poyhonen (1963) were the first to apply Newton law of gravitation in examining trade flows. They assumed that, as planets are commonly pulled in the extent to their sizes and proximity, so do nations in the extent to their GDP and proximity (Becchetta et al., 2012). Their gravity theory incorporated economies of scale by accounting for 28 University of Ghana http://ugspace.ug.edu.gh market size, proxied by nation population size and GDP. A geographical dimension in the form of distance between trading partners was incorporated into the model. Consequently, the gravity model of trade assumes that bilateral trade between two nations is positively proportional to their economic size (GDP) and negatively with the distance between them. The multiplicative form of the basic gravity model of trade is presented below: = (3.2) Where is the export flows between country i and j, represents the GDP of country i and j respectively, measures distance between the capital cities of country i and j, and finally s are coefficients of the variable noted above. The log-linear form of equation (3.2) is shown below: = + + + (3.3) Where and are expected to be positive whereas is expected to be negative as distance is a proxy for transport cost and its interpretation will means that, as distance between trading roots increases, the cost of transportation will rise. Although, the model has witnessed wide usage in empirical studies in recent time, it came under serious criticism for lacking theoretical basis. 3.2.1. Theoretical Justification for the Gravity Equation Since the basic gravity model of trade came under serious criticisms between the 1960s and 1970s for lacking theoretical basis, it casted aspersion on the respectability of the empirical predictability power (Frankel, 1997). However, with the increasing influence of geographical and macroeconomic elements in bilateral trade flows, the model began to re-attract interest again from the late 1970s to give theoretical clarifications to such trade issues. The theoretical validity of the gravity equation was greatly enhanced by the works of Anderson & Wincoop (2003), 29 University of Ghana http://ugspace.ug.edu.gh Deardorff (1995), Helpman & Krugman (1985), Bergstrand (1985, 1989) and Anderson (1979). Their works demonstrated that individual trade models such as the Ricardian theory, Heckscher Ohlin model and others can be derived through the gravity model of trade. Firstly, the work of Linnerman (1966) was the earliest endeavour made to give a hypothetical premise to the gravity model of trade by including population variable to the model, but, it was the work of Anderson (1979) that provided the most critical attempt in upgrading the theoretical premise of the gravity model based on the Armington Assumption. His work applies product differentiation and adopted a Cobb-Douglas preference, and postulated that products are differentiated by their countries of origin. Anderson (1979) used the linear expenditure system in which every good is produced by just a single nation and nation’s preferences for a good are thought to be homothetic, and uniform across importing nations which is approximated by constant elasticity of substitution (CES) utility function. Without considering tariffs and transport cost, Anderson stated that his mode of utilization of the gravity model is an alternative to the cross-sectional budget studies. However, Kristjansdottir (2002) pointed out that Anderson’s model is restricted by the fact that it is only valid for nations with identical preferences for a traded products and identical structure as far as transport cost and trade tax are concerned. Similar to Anderson (1979), Bergstrand (1985) applied the Armington Assumption and assumed constant elasticity of substitution preferences and a microeconomic foundation of a simple monopolistic competition model to explain his gravity model of trade. He tested his model and concluded that price levels and exchange rate have a reasonable and significant impact on trade 30 University of Ghana http://ugspace.ug.edu.gh flows. As noted by Rahman and Dutta (2012), Bergsrand called his model the generalized gravity model of trade. The idea from the work of Helpman and Krugman (1985) that bilateral trade relies upon the product of the trading partners GNPs, provided a further theoretical foundation to the gravity model of trade. As indicated by Frankel (1997), the approach utilized by Helpman and Krugman (1985) is the one we build upon if the intended interest is to examine the theoretical effect of the preferential trade agreement on the volume of trade and economic welfare of the people. The work of Deardorff (1995) contributed to the theoretical foundation of the gravity model by deriving the gravity model of trade from the framework of the Heckscher-Ohlin model. His model is just a simplification of Anderson’s (1979) model and the presumption that same preferences hold not for a specified trade good like Anderson (1979), but for all traded goods. He stated that with the absence of trade barriers, trade in homogeneous products makes producers and consumers indifferent toward trading partners both domestic and international trade, so long as they sell or buy the desired goods. Makochekanwa and Jordaan (2008) pointed out that Deardorrf (1995) assumption influenced him to derive an expected trade flows that correspond exactly to a simple frictionless gravity model whenever preferential are identical between trading partners. The works cited above gave a sound clarification to the theoretical premise of the gravity model of trade, but, none of them showed clearly the part that distance plays on trade. The famous article of Anderson and Wincoop (2003), "Gravity with Gravitas: A solution to the Border 31 University of Ghana http://ugspace.ug.edu.gh Puzzle" provided an exceptionally key effect on the theoretical foundation of the gravity model of trade. The authors were displeased with the theoretical backing of the gravity model and added a remoteness variable termed as a Multilateral Trade Resistance (MTR). Their model is considered one of the most essential contributions to the theoretical establishment of the gravity model of trade. However, Anderson and Wincoop (2003) model came under criticism on the grounds of the postulation made that all traded products are differentiated by the country of origin and that each country has to specialize in the production of only one simple product. Furthermore, the assumption made was not realistic, hence, unable to provide a concrete formula for estimation of the CES between the traded products, but rather assumes a value for it. It was the works of Bergstrand et al., (2013) that provided several ways of estimating the CES. In conclusion, the literature cited above indicates that the theoretical foundation and empirical success of the gravity model of trade had been proven, hence, justifying its usage. 3.3 Empirical Review Determinants of exports performance have received widespread attention from policy makers and researchers around the world because of their vital role in the economic growth of many countries. Hence, knowing and understanding the mechanism through which various economic and policy drivers influence exports is key to the implementation of appropriate trade policy measures to further enhance exports and shield economies from potential adverse implications of developments from external markets. Various internal and external factors have been studied in the exports literature as potential determinants of exports outcomes and sometimes contradictory findings regarding the effects of such factors on exports based on demand and supply. This 32 University of Ghana http://ugspace.ug.edu.gh section will first review works that adopted the gravity model of trade to analyze the determinants of exports growth, and then review studies that have adopted other approaches. 3.3.1 The Gravity Model of Trade and Exports Performance Given the vast utilization of the gravity model of trade as an empirical strategy in examining pattern and determinants of exports and imports flows between countries, the target of this section is to review some previous studies as a guide in choosing the model and factors to be considered in the study. Timbergen (1962) was the first economist to utilize the gravity model in econometrics to examine the determinants of international trade flows. Using data including 18 nations, his findings revealed that income and distance are statistically significant in determining trade flows. The study further revealed that adjacency and being a member of Britain Commonwealth and Benelux FTA all play a vital role in trade flows. Bergstrand (1985) adopted the gravity model of trade to examine factors that determine bilateral exports flows across 15 OECD nations. The aim of the study was to address issues confronting the theoretical foundation of the gravity model of trade. Hence, in addition to the basic gravity model of trade variables, he incorporated variables such as exchange rate, imports and exports price indices, GDP deflator of both countries, and dummy variables like adjacency, European Economic Community (EEC), and European Free Trade Area (EFTA) as explanatory variables. The result of the study revealed that economic sizes of both countries, imports price index, adjacency and EFTA membership have a significant positive influence on bilateral exports 33 University of Ghana http://ugspace.ug.edu.gh flows, whereas variable such as the geographical distance between the capital cities of the trading partners was found to negatively affect the volume of trade between countries. The remaining variables used in the estimation were found to be statistically insignificant. Since cross-sectional data was used in the estimation, the outcome of the study suffers from the problem of heteroscedasticity. Thursby and Thursby (1987) examined the effect of exchange rate instability on bilateral trade flows: A gravity model of trade approach. Applying panel data of 17 countries from 1974-1984, the findings of the study revealed that coefficients of variables such as distance and adjacency are statistically significant and with their expected signs. Furthermore, the exchange rate was negative and significant. Finally, the preferential dummy appeared to be more important for European Free Trade Arrangement (EFTA) countries than for those in the European Economic Community (EEC). To avoid heteroscedasticity problem, the adopted an estimation method that retains the least squares coefficients but uses a covariance matrix estimator that is consistent. The limitation of this study is that the authors failed to address the issue of autocorrelation. Another study by Oguledo and Macphee (1994) estimated a reformulated gravity model of trade derived from the linear expenditure framework with trade flows data from 162 countries for 1976. Their study revealed that preferential trade arrangement and price variables were significant with their expected signs. Their findings also confirmed that other factors such as GDP, population and distance which are commonly used in the gravity model have significant influence on trade flows. Since the authors used cross-sectional data for the estimation, the outcome of the study is affected by the problem of heteroscedasticity. 34 University of Ghana http://ugspace.ug.edu.gh Osei-Assibey K. (2017) investigated how exchange rate instability, earnings uncertainty affect the bidirectional trade flows in Ghana. He employed the augmented gravity model of trade and the fixed effect pooled cross-sectional estimation by using OLS robust errors on three regressions such as bilateral trade, exports and imports. The estimation of the study revealed that exchange rate instability is negative and statistically insignificant under bilateral trade and imports results while positive and significant under exports result. The outcomes on the income and per capita income were found to be positively and negatively related to three regressions estimated respectively and both variables were statistically significant under both regressions. Braha et al., (2016) examined the determinants of Albanian Agricultural Exports. The study employed the conventional gravity model covering the exports of the Albanian economy from 1996-2013 and the Poisson Pseudo Maximum Likelihood (PPML) estimation technique. Their empirical findings revealed that export flows increase with increasing domestic economic size (GDP). Furthermore, the results showed that the distance variable has a negative effect on the exports of agricultural products whereas variables such as sharing common border and price stability (inflation), RTAs and institutional quality of the trading partners facilitate Albanian agricultural exports. Similarly, Dlamini et al., (2016) examined factors that determine Swaziland sugar exports using the gravity model of trade. Applying a panel dataset of Swaziland and its 25 major trading partners from 2001- 2013, the findings of the study uncovered that GDP of Swaziland and her trading partners, importers land size areas, official common language, the creation of COMESA, 35 University of Ghana http://ugspace.ug.edu.gh and EU trading blocs have a significant positive effect on Swaziland sugar exports. On the other hand, the study revealed that the population of the trading partner, trade openness of Swaziland and distances have a significant negative impact on the exports of sugar from Swaziland. Jordaan and Eita (2011) carried out an empirical study to investigate the determinants of South Africa’s wood exports using the gravity model approach. The study applied a panel data on South Africa and its 68 trading partners from 1997-2004. They revealed that South Africa’s population and importer's GDP are positively related to wood exports with the coefficients being statistically significant. The population of the importer and GDP of the domestic country affect exports of wood products negatively. The finding on importer’s population indicates that the trading partner becomes self-sufficient as their population grows. However, the distance variable was statistically insignificant. The results on regional trade arrangements were found to be very influential on wood exports. For example, membership of NAFTA and EU reflected a negative impact on exports while that of the South African Development Committee (SADC) increases exports of wood products. Finally, the results revealed that unexploited trade potential exists from 2002 to 2004 between South Africa and its trading partners over the study period. Using the gravity model of trade, Habab et al (2010) investigated the determinants of Egyptian agricultural exports growth. The study utilized panel data on Egypt and her 50 major trading partners from 1994 to 2008. Their estimated results revealed that Egypt’s GDP, exchange rate volatility has a positive effect on Egyptian agricultural trade whereas GDP per capita and transportation costs, proxied by distance influence it negatively. All these were statistically significant. 36 University of Ghana http://ugspace.ug.edu.gh Chan et al., (2007) applied the gravity model of trade to examine the determinants of China’s textiles exports. Their study applied panel dataset on China and her top 10 importing countries from 1985 to 2004. The findings of the study indicated that GDP of China and her importers GDP, GDP per capita of the importer, being a common membership in a trade organization affect exports positively whiles population growth rate, real bilateral exchange rate have a negative impact on China’s textiles exports. All the variables were statistically significant except the distance variable. Kristjansdottir (2002) adopted the gravity equation to examine determinants of Icelandic exports growth, using panel dataset on Iceland exports based on four sectors and its 16 top trading partners over a period of 11 years. Her empirical findings showed that the size and income of Iceland do not seem to matter much on the volume of exports, not even when corrected for the small country sizes but regional trading blocs and individual sectors matter. Finally, the study also revealed that marine products varied considerably in their sensitivity to distance and country factors. 3.2.2 Other models and exports performance Due to the limited availability of literature on gravity model of trade and exports (specifically non-traditional exports) performance in Ghana, this section will review literature that has employed other econometric models to analyze the subjects in Ghana and the rest of the world that are significant to the study. 37 University of Ghana http://ugspace.ug.edu.gh Anagaw and Demissie (2013) adopted the Johansson Co-integration and Vector Error Correction technique to examine the determinants of exports performance in Ethiopia from 1970 to 2011. Their empirical findings showed that the home country’s GDP, real effective exchange rate, financial development, trade liberalization and infrastructural development influence exports performance of Ethiopia positively and statistically significant. Real GDP of the trading partners were found to be statistically insignificant in determining exports performance. Skosana and Kabuyab (2014) applied Cointegration technique to examine the determinants of Swaziland’s exports performance from 1980 to 2010. The study tested for Cointegration and vector error correction to establish the long run relationship between the variables and to capture the speed of adjustment to the long-run equilibrium. Findings from their study revealed that inflows of foreign direct investment, world demand and nominal exchange rate are key significant factors in determining exports performance in the long-run with elasticities of 0.58, 1.19 and 0.36 respectively whereas the short run elasticities were 0.26, 4.3 and 0.50 respectively. Boansi et al., (2014) examined factors that enhance agricultural exports trade of fresh pineapples from Ghana. Using detached regression equation of exports value and volume as the dependent variables, the study applied the Ordinary Least Square Estimator. The findings of the study revealed that exports of fresh pineapples from Ghana have a comparative advantage and are more value-driven than volume driven. Furthermore, the study showed that production level, trade openness and competitiveness index of Ghana have a positive impact on both value and volume of pineapple exports from Ghana. However, the domestic demand and net foreign direct 38 University of Ghana http://ugspace.ug.edu.gh investment inflows have a negative effect on fresh pineapple exports. All these variables were statistically significant Using Tobit estimation approach, Agyei-Sasu et al., (2010) examined the intensity of exports success of the horticultural enterprises in Ghana. They used primary data gathered from 52 managers and representatives of horticultural exporting firms in Ghana. Their empirical findings indicated that manager’s educational level, manager’s job experience, manager’s training in exports management, manager’s entrepreneurial orientation, the presence of exports department, product diversification and government support directly influence the intensity of exports success. On the other hand, exports barriers and constraints in accessing working capital negatively impact the intensity of exports success in Ghana. Kyereboah-Coleman et al., (2006) applied the generalized least square (GLS) estimation technique to study the link between cooperate governance and non-traditional exports performance for Ghana using panel dataset from 1995-2004. The study showed that indigenous ownership of NTEs firms and non-executive director’s inclusions on their boards ensure efficiency in NTEs performance in Ghana. 3.3. Conclusion As pointed out earlier, the focal objective of this chapter was to review previous empirical studies to serve as a guide in the selection of the appropriate econometric model and variables to be used in our study. In the next chapter of this study, we will discuss the research methodology applied in this study. 39 University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR METHODOLOGY 4.0 Introduction The focal aim of this chapter is to present the empirical technique to be adopted toward the attainment of the study objective. The chapter is divided into five main sections. The first section of the chapter presents the specification of the gravity equation adopted for the study; the second part focuses on the description of variables used in the study, the measurements and sources from which we obtained them; the third section provides description of the data used: the fourth part presents issues relating to the econometric methodology including the estimation technique and robustness checks, and the final section presents the measurement of Non-Traditional Exports (NTEs) potential of Ghana. 4.1. Model Specification For the purpose of examining the determinant of Non-Traditional Exports (NTEs) growth in Ghana, the study adopted the augmented gravity equation of trade. The motive behind the use of the gravity equation is that it has been widely used to examine factors that determine bilateral trade flows due to its empirical robustness and explanatory power capabilities. According to Kepaptsoglou et al. (2010), the gravity model provides an excellent empirical robustness outcome when analyzing international trade flows. Since the motive of this study is to investigate factors that affect NTEs growth in Ghana, we adopted a modified gravity model of trade employed by McCallum (1995), and later used by Braha et al., (2016). The logarithmic form of the modified model used by McCallum is presented below: 40 University of Ghana http://ugspace.ug.edu.gh = + + + + + (4.1) Where TRAij represent trade flow from country i to j; GDP represents the Gross Domestic Product ; Dij indicates the geographical distance between the capital cities of country i and j; DUMij is a dummy variable that determines trade flows; Uij is the stochastic error term, βn and λk are the parameters of the model to be estimated. Equation (4.1.) is expanded by adding several variables to fit appropriately the gravity model for the Non-Traditional Exports in Ghana. Here we adjusted the modified form of the gravity equation and presented the log-linear of the desired model to be applied in the estimation as follows: = + + + + + + + + + + + + + + + + + (4.2) Where the dependent variable EXPgjt represents exports of non-traditional product from country g (Ghana) to country j at time t. The explanatory variables are Gross Domestic Product (GDP), Population size, Distance between capital cities of country g (Ghana) and j (DISS), the land area of the trading partner (AREAj), real bilateral exchange rate between country g and j (RBERgjt), Openness of trade of the trading partner (OPENNESSgjt), institutional variables such as political stability and absence of violence index (POLIgjt), rule of law index (RLgjt), and the regulatory quality index (RQgjt) ) of the trading partner, variables such as EUgj, ECOWASgj, LANDLOCKgj, BORDERgj, COMLANgj and ASEANgj are dummy variables representing European Union trading bloc, Economic Community of West African State trading bloc, landlocked nature of the trading partner, whether country g share border with j, whether country 41 University of Ghana http://ugspace.ug.edu.gh g share common official language with j and Association of Southeast Asian Nations trading bloc respectively and εij is stochastic disturbance term that is assumed to be well-behaved. 4.2. Description of Variables and Data Sourced 4.2. 1 Dependent variable Non-Traditional Exports (EXPgjt) of Ghana is used as the dependent variable of this study. It measures the foreign income received annually from exports of Non-Traditional products ( ). Thus exports ( ) refers to the total value of all Non-Traditional products flowing out from Ghana to a given trading partner valued in U.S. dollars. Data on NTEs were obtained from Ghana Exports Promotion Council (GEPC) database. 4.2.2 Independent variables Gross Domestic Product ( and ): The variables and are used to indicate economic size of Ghana and her trading partners respectively. Thus, measures productive capacity of Ghana whereas measures absorptive capacity of the importing countries. This implies that a large market size of importing country should impact positively on the demand for Ghana’s NTEs. Since the GDP of Ghana measures the productive capacity of the country, an increase in it will affect the supply of NTEs positively. Hence both variables are expected to influence NTEs growth positively. Information on GDP for both countries were acquired from World Bank Development Indicators (WDI) database and valued at current US dollar. Population ( and ): The population variables measure the total population of Ghana and her trading partners. The variable represents the population of Ghana and theoretically, it is expected to influence NTEs positively, if Ghanaian population is not a net 42 University of Ghana http://ugspace.ug.edu.gh consumer of non- traditional products but rather net producer. On the other hand, the represent the trading partner population and it is expected to influence NTEs positively if the trading partner’s population is a net consumer of NTEs from Ghana. Hence, the expected sign for and are ambiguous, since, it depends on the consumption or production nature of both population. The data for these variables were retrieved from World Bank Development Indicators (WDI) database. Distance ( ): The distance variable measures the geographical distance between the capital cities of Ghana and her trading partners, measured in kilometers (km). The distance variable is expected to be negative since it serves as substitution for the transportation costs between Ghana and the importing country. Data on the distance variable were sourced from World Distance Calculator online database measured in kilometers. Land Area of the importing country ( ): This variable measures the physical land size in square kilometers of the trading partner. It is expected to influence NTEs negatively, since, a large land area of the importing country presents her the opportunity to expand it production capacity, hence decreasing demand for foreign products (NTEs). Data on this variable were retrieved from World Fact-book online database. Real bilateral exchange rate ( ): This variable measures the real exchange rate that exists between the Ghana cedi and the trading partner currency. This variable represents the price of Ghana cedi expressed in terms of foreign currency of each trading partner. We calculated this variable by multiplying the nominal exchange rate of each trading partner ( ) via the ratio of 43 University of Ghana http://ugspace.ug.edu.gh the trading partner’s price index ( ) to domestic price index ( ). The formula is presented below: = The real bilateral exchange rate variable is projected to influence NTEs growth positively, since; an increase in this variable indicates that less foreign currency of the trading partners can be exchanged for more cedis. The data on nominal exchange rate, trading partner price index and domestic price index were gathered from IMF statistics database. Institutional quality: The institutional variables such as political stability and absence of violence index ( ), rule of law ( ) and regulatory quality ( ) measures the institutional excellence that Ghana’s NTEs are confronted with at the capital cities of the trading partners. Variables such as political stability and absence of index ( ) and rule of law ( ) are expected to influence NTEs positively, since, an improvement in such variables provide security for Ghana’s NTEs. On the other hand, an improvement in institutional variable such as regulatory control quality ( ) is expected to influence NTEs from Ghana negatively, if Ghana NTEs do not meet the standard of quality checks from the importing country. Institutional quality variables were sourced from World Bank Governance Indicator online database. Trade openness ( ): The openness index variable indicates the level of trade openness in the importing country. It is calculated by adding exports and imports of the trading partner and dividing the result by the trading partner’s GDP. The formula is presented below: = 44 University of Ghana http://ugspace.ug.edu.gh Where represent exports, imports and gross domestic products respectively of the importing country. Openness variable is expected to have a positive effect on Ghana’s NTEs growth, since, the more open countries are to trade, the lesser their trade restrictions, hence, the lower the cost of exporting products to these countries. Data on variables were gathered from World Bank Development Indicators (WDI) database. The Economic Community of West African State ( ): This is a dummy variable representing regional trading bloc inside the West Africa States, and takes the value of 1, if the importing country is a member and 0 if not. The effect of on the Ghana’s NTEs is projected to be positive, since; it presents member states with larger market opportunity and trade policy such as the ECOWAS trade liberalization Scheme which enable members to enjoy duty free. Hence, exporting Ghana non-traditional products to such ECOWAS reduces cost associated with exports. Data on ECOWAS were attained from the community’s website. European Union ( ): This dummy variable represents regional trading bloc from the European countries, and it takes a value 1, if the importing country is a member, but 0 if not. This variable is anticipated to influence Ghana’s NTEs growth positively, since the Union presents Ghana’s NTEs with enormous marketing opportunity and trade policy such as Economic Partnership Agreement between the Union and ECOWAS which enables members to trade under minimum trade restrictions. Data on European Union were sourced from the union’s website. 45 University of Ghana http://ugspace.ug.edu.gh Association of Southeast Asian Nations ( ): is a regional trading bloc dummy that is equal to 1 if the trading partner is a member and 0 otherwise. It is expected to influence NTEs growth positively, since; it presents Ghana’s NTEs with enormous marketing opportunity. Data on Association of Southeast Asian Nations were sourced from the Association’s website. Other bilateral dummy variables added to the model are which takes the value of 1 if Ghana and her trading partner share a common land border and 0 if not, is a dummy variable that takes a value of 1, if Ghana and the trading partner share common official language (English) and 0 if not, and is dummy variable that measures the non-coastal nature of the importing country and assumes the value of 1 if the country is a landlocked country and 0 if not. Dummy variables such as sharing a common border ( ), and common official language ( ) are expected to facilitate trade, since, they reduce the cost associated with trade. Dummy variable like is expected to influence NTEs negatively, since; it makes the cost of transportation very high. Data on these variables were sourced from the World Fact Book website. 4.3. Data Description As noted by Gomez and Milgram (2010), the common practice with respect to the estimation of the gravity model before 1990s was the employment of cross-sectional data, but estimation through cross-sectional data does not control for heterogeneity among country and consequently could lead to estimation bias. Hence, to lessen the predisposition related with assessing the gravity model of trade with the utilization of cross-sectional data, works by Matyas (1998), Egger (2000), Rose and Wincoop (2001), Egger and Pfaffemayr (2004), Glick and Rose (2002), 46 University of Ghana http://ugspace.ug.edu.gh Melitz (2007) moved in the direction of panel data that is cross-section of several consecutive years. The uses of panel data have proven by researchers to be advantageous over the use of cross-sectional and time series data. As noted by Gomez and Milgram (2010), panel data enable the study of the characteristics of variables over time and help to avoid the potential problem of multicollinearity that is associated with cross-sectional data. Given these numerous benefits of panel data, this study employs a panel dataset of annual observations on 78 major trading partners of Ghana over a period of 13 years (between 2004 and 2016 ) to examine the determinants of Non-Traditional Exports (NTEs) growth in Ghana within the framework of the gravity equation. Report from GEPA (2016) shows that Non-Traditional products are exported to about 137 countries but the 78 countries selected imported about 98% of the total NTEs from Ghana during the period under study. The decision regarding the sample period and the countries involved in the study are influenced by the relative significance of each trading country and the availability of data on every one of the variables utilized in the study. In sum, the annual data covers 78 major trading countries for the year 2004 to 2016 with one dependent variable and 17 explanatory variables ( thus, an aggregate of n = 1003, N = 78, and T = 13). 4.4 Estimation Techniques 4.4.1 Log Linearization model Until the early 2000s, the traditional gravity equation estimation technique was to transform the model in log-linear form as shown in equation (4.2) and apply panel estimation model like pooled regression, random and fixed effect model. 47 University of Ghana http://ugspace.ug.edu.gh 4.4.1.1 Pooled regression model The pooled regression model is a type of panel estimation method in which the estimation overlooks the panel structure of the data. The advantage of utilizing this method is that it can clarify the past while simultaneously forecasting the future characteristics of exogenous factors in connection to endogenous factors (Moon et al., 2004). The principal deficiency of this model is that it doesn't take into account the estimation of country-specific effect and accept that all countries are homogenous. The model can be seen as a basic model; however, it is extremely restrictive among alternative models. 4.4.1.2 Random effect model This model assumes that the individual (heterogeneity) effects are taken by the intercept and a random part ( ) which is autonomously and indistinguishably distributed over the individual countries. One advantage of this model is that, it permits the estimation of the impact of time invariant factors which are cancelled out in fixed effect model. This model is suitable when evaluating the flow of trade between random samples drawn from large group of trading countries (Habab et al., 2010). 4.4.1.2 Fixed effect model The fixed effects model is a model that takes into consideration the fact that each cross-sectional unit might have some unique attributes of its own. In this estimation technique, the intercept in the regression is allowed to differ among individual units, but constant within a unit over time. The fundamental weakness of this technique over the others is that it doesn't consider estimation of the coefficients of time-invariant factors. This model is suitable when evaluating trade flows between ex-ante predetermined choices of trading partners (Habab et al., 2010). 48 University of Ghana http://ugspace.ug.edu.gh The main distinction between fixed and random effects as noted by Greene (2008) is whether the unobserved individual effect embodies components that are correlated with the regressors in the model and not whether these effects are stochastic or not. Hausman test is used to test whether the fixed effects and random effects estimators are essentially unique from each other. The Hausman test is performed based on the null hypothesis that the two estimation techniques are both consistent, and that, they should produce outcomes that are similar. Hence, rejection of the null hypothesis implies that the fixed effect model is the appropriate estimation method over the random effects model. 4.4.2 Non-Linear Models Although, the above log-linearization estimation technique has been applied in several literatures to estimate the determinants of bilateral trade flows, this technique for estimating the gravity model of trade came under serious criticism in the early 2000s. Silva and Tenreyro (2006) highlighted the basic issues concerning the log-linear models that have been overlooked in theoretical and applied studies. Specifically, their contention was that the logarithmic transformation of the gravity model like equation (4.2) is not a relevant approach in the estimation of the elasticities. They pointed out that, the log-linearization of the empirical model in the presence of heteroscedasticity prompts inconsistent appraisals in light of the fact that the expected value of the logarithm of the random factors relies upon a higher moment of its distribution. Thus in the standard gravity model of trade presented below: = + + + + (4.3) Where the expected value of the log-linearized equation (4.3) will be: E = E (4.4) 49 University of Ghana http://ugspace.ug.edu.gh E = E + E + E + E + E (4.5) Since from Jensen’s inequality, the E ≠ E is an indication that the conditional distribution of is altered, hence, evaluating the standard gravity model of trade through OLS approach will bring about misleading of estimates. Although, the heteroscedasticity issue won't influence the parameters estimated, it will bias the variance of the estimated parameters, and hence, the t values for the estimated coefficients can't be trusted (Gomez and Milgram, 2010). Furthermore, Silva and Tenreyro (2006) pointed out that the transformation of the standard gravity model of trade into log linearized model under a dependent variable ( ) with zero values will create additional problem in estimation of the parameters. Given the issue of heteroscedasticity and zero trade values confronting the log-linearization of the standard gravity model of trade, Silva and Tenreyro (2006) proposed a Poisson Pseudo Maximum Likelihood (PPML) estimator. The authors stated that the dependent variable of the gravity model ought to be evaluated in levels as against applying logarithms; hence, the issue of heteroscedasticity and zero values will be avoided. As indicated by Silva and Tenreyro (2006), the PPML is the appropriate model since it accounts for observed heterogeneity, gives natural way to deal with zero trade values and finally, it avoid under-prediction of extensive trade volumes and flows by producing estimates of trade flows and not the log of trade flows as in case of log-linear estimation. The numerical estimator of the PPML is presented below: =0 (4.6) Where Silva and Tenreyro (2006) claim that when the conditional mean (E = exp ) is well specified makes the estimator in equation (4.5) consistent. 50 University of Ghana http://ugspace.ug.edu.gh Although, the PPML estimator has been used in several literatures such as Martin and Pham (2008), Burger et al., (2009), Liu (2009), Westerlund & Wilhelmsson (2011), Covaci & Moldovan (2015), Martinez-Zarzoso (2013), Braha et al., (2016) and so on, it has come under criticisms. As indicated by Martin and Pham (2008), the PPML estimator is less subjected to bias coming about because of heteroscedasticity, however not jointly proven to be robust under heteroscedasticity and zero trade flows issues. Furthermore, Burger et al., (2009) pointed out that the standard Poisson model is sensitive to the issues of over-dispersion and excess zero trade flows, hence will not be an appropriate estimator for trade values that contain zeros. To solve the challenges that confronted PPML estimator, several non-linear models have been proposed. Martinez-Zarzoso et al., (2007) proposed the Feasible Generalized Least Squares (FGLS) as the best estimation system if the exact form of heteroscedasticity in the data is overlooked, since, it weighs the observation as per the square root of their variance and robust to any form of heteroscedasticity. Manning and Mullahy (2001) proposed the Gamma Pseudo Maximum Likelihood (GPML) estimator. The authors stated that the conditional variance of the dependent variable under their model is assumed to be proportional to its conditional mean. Burger et al., (2009) proposed the utilization of the Negative Binomial Pseudo Maximum Likelihood (NBPML) to take care of the issue of over-dispersion confronting the PPML estimator. Helpman et al., (2008) used the Hackman sample selection model and altered it to consider the bias connected with the heterogeneity. However, Liu (2009) contends that the Hackman gravity model embraces the log-linear specification as the normal OLS; hence, it is subject to the issue of heteroscedasticity raised by Silva and Tenreyro (2006). Because of this criticism, Silva and Tenreyro (2011) expressed that other models may outperform the PPML, yet still, the PPML 51 University of Ghana http://ugspace.ug.edu.gh ought to be the benchmark against which other alternative estimators ought to be compared because of its identified advantages. Silva and Tenreyro (2011) and Soren and Bruemmer (2012) further stated that, the PPML is reliable and well-behaved even within the sight of over- dispersion in the variables (i.e. at the point when the conditional variance isn't equivalent to the conditional mean) and that the predominance of large proportion of zeros does affect its performance. In sum, the estimation techniques mentioned above show that each model has advantages and disadvantages; hence, it can't be asserted that any of the models absolutely outperforms the others. So for the purpose of the study and the nature of the data (less zero values), we adopted an econometric approach using the PPML estimator and the log-linear estimation models such as Pooled regression, Fixed effects and Random effects. As stated by Gomez (2013), due to the current misunderstanding concerning the best estimation techniques, it has become a frequent practice in literature to add more than one estimation method for the same dataset and apply post-estimation test to decide which model to settle on. The study will estimate equation (4.2) with log-linear OLS approach whereas equation (4.6), which is the transformation of equation (4.2) in levels will be estimated using PPML. Equation (4.6) is presented below: = + + + + + + + + + + + + + + + + + (4.6) 52 University of Ghana http://ugspace.ug.edu.gh 4.5 Robustness check The study will mainly perform the park type test, Gauss-Newton Regression test and the RESET test to aid in deciding whether the PPML estimator is better than the estimator under log-linear model (OLS) for estimating our gravity equation for the determinants of NTEs growth in Ghana. 4.5.1 The Park type test The park test is performed to justify the utilization of the non-parametric estimator of the variance and examine whether the extent of heteroscedasticity pattern assumed by the model is fitted correctly. From Manning and Mullahy (2001), if V = (4.7) holds, and E can be estimated consistently, then can be estimated through a Park type auxiliary regression. Park (1966) auxiliary regression model is presented below: = + + (4.8) Where is the estimated value of E . The test permits studying on the hypothesis that constant elasticity model can be evaluated in the log linear model. Thus, if the hypothesis of constant elasticity is rejected, it indicates that estimation in the log linear form isn't appropriate for estimating the gravity equation. Therefore, estimating the gravity model through PPML estimator is more appropriate. 4.5.2 Gauss-Newton regression test The Gauss-Newton regression test (GNR) is tested on the hypothesis that V is proportional to E and it is rejected, if the confidence interval for in equation (4.9) below does not include 1 (Silva and Tenreyro, 2006). = + (4.9) 53 University of Ghana http://ugspace.ug.edu.gh Where =1 and = exp the equation (4.9) could be expanded to as follow: = + + (4.10) Where the hypothesis that V is proportional to E is tested by verifying if whether is statistically significant. Due to the possibility that the error term in equation (4.10) will be heteroskedastic, Gauss-Newton regression is perform through the use of Weighted Least Squares method. The Weighted Least Squares method of the Gauss- Newton regression model is presented below: = + + (4.11) Where the GNR test is performed by estimating equation (4.11) by OLS and concluding on the statistical significance of . Hence, obtaining statistically insignificant of implies that PPML assumption V = E cannot be rejected, thus justifying the use of the PPML estimator. 4.5.3 Regression Equation Specification Error Test (RESET) The RESET test is a test that enables us to confirm whether the functional form of the estimated gravity equation is effectively defined. Thus, if the functional relationship between the reliant variable and the explanatory variables is non-linear in nature, then the estimated model is said to be mis-specified. Therefore, testing the statistical significance of the additional variable, we can confirm the appropriateness of the estimated model (Sillva and Tenreyro, 2006). 54 University of Ghana http://ugspace.ug.edu.gh 4.5 Predicting Unexhausted Destination and the Speed of Convergence for Non- Traditional Exports (NTEs) Literature over the past years has depended on two main ways to compute exports potential (unexhausted exports), which are, out-sample and in-sample approach. Given the purpose of this study, we adopted the out-sample approach to predict unexhausted trade destination available for Ghana’s NTEs as specified in equation (4.6). This approach involves the utilization of the estimated parameters to project the natural trade relation between the importing nations so that the difference between the actual exports and the predicted exports represents the un-exhausted exports destinations (Wang and Winters, 1992; Hamilton and Winters, 1992; and Brulhart and Kelly, 1999). As noted by Kaur and Nanda (2011), a positive unexhausted export indicates future possibilities of exports expansion while a negative value demonstrates that the exporting nation has exceeded its exports potential with a specific trading partner. The study will estimate exports potential (unexhausted) by calculating the difference between the potential exports (P) and actual exports (A). The formula is presented below: (4.12) Where is an unexhausted NTEs of Ghana at time t whereas and represents the predicted and actual NTEs values respectively. As noted above, empirical studies often use equation (4.12) to estimate trade potential between trading partners. Egger (2002) criticized the method based on econometric mis-specification; hence, Jakab et al. (2001) suggested the addition of speed of convergence when predicting the potential between trading partners. The speed of convergence is defined as the average growth 55 University of Ghana http://ugspace.ug.edu.gh rate of potential exports divided by the average growth rate of actual exports between the years under study. Due to issues of zero export values in some years, the study will calculate the speed of convergence through the use of average instead of average growth as defined above. The formula for the speed of convergence for this study is presented below: -100] (4.13) From equation (4.13), there is a speed of convergence if the average exports potential is lower than that of the actual exports and the estimated speed of convergence is negative. On the other hand, there is a speed of divergence if the average exports potential is higher than that of the actual exports and the estimated value is positive. 56 University of Ghana http://ugspace.ug.edu.gh CHAPTER FIVE ESTIMATION AND DISCUSSION OF RESULTS 5.0 Introduction The focal aim of this chapter is to present and discuss matters associated with the econometric estimation of the model specified in chapter four of this study. The study adopted STATA in estimating the regression models mention in Chapter Four of this study. The chapter starts by conducting a descriptive statistical analysis of the non-binary variables utilized as part of the study. We then proceed to perform robustness checks discussed in the previous chapter. This is followed by the presentation and discussion of the results of the preferred estimation model. The final part of the chapter presents the prediction of un-exhausted destinations for NTEs of Ghana. 5.1 Descriptive Statistics and Correlation Matrix This section quickly analyzes the essential statistical properties and correlation of the non-binary variables included in the model of the study. The correlation matrix in Appendix 1.7 shows that the correlation coefficients between the dependent export variable and the corresponding explanatory variables relatively low, hence the absence of the endogeneity problem. The summary statistics included in the study are the mean, standard deviation, minimum and maximum values. This information is presented in Table 5 -1 below. From Table 5 -1, total NTEs from Ghana within the study period average about US$ 22.1 million. The minimum and maximum values of NTEs from are US$0.0 and US$ 403 million respectively. The average GDP of Ghana and its trading partners for the period was valued about US$ 34.5 billion and US$ 782 billion respectively. By comparison, the average GDP of Ghana and its trading partners indicate 57 University of Ghana http://ugspace.ug.edu.gh that the trading partner is economically larger than Ghana. Also, Ghana NTEs travels an average distance of 5569.24 km to its market destination. Finally, the variable with the larger standard deviation is GDP of the importing country. Table 5- 1: Summary Statistics of the Dependent and Independent Variables within the Period under Study. Variables Mean Standard Deviation Minimum Maximum EXPORTgjt 2.21e+07 4.39e+07 0.0000 4.03E+08 GDPgt 3.45e+10 8.94e+09 2.22E+10 4.82E+10 GDPjt 7.82e+11 2.03e+12 6.90E+08 1.69E+13 POPgt 2.46e+07 2251414 2.10E+07 2.82E+07 POPjgt 7.13e+07 2.06e+08 322526 1.38E+09 DISTANCEgj 5569.24 3605.662 169.14 15459.05 RBERgjt 0.5472735 0 .7040116 0.0000 3.97657 RQjgt 58.59121 28.9801 2.4630 100 RLjgt 56.24159 30.17516 0.3121 100 Opennessjgt .9403621 0.6815336 0.0000 4.4262 AREAj 1261535 2862108 719.20 1.71E+07 POLSTAjgt 48.27007 28.12483 0.4739 100 Source: Author’s Computation using STATA 5.2 Diagnostic and Robustness checks To determine the appropriate estimation technique for the analysis of the specified model, the study performed tests such as time effect, heterogeneity, heteroscedasticity, Park-type test, Gauss-Newton regression test and the RESET test. 5.2.1 The Time Effect Test In testing for the time effect, the study introduces new variables called the year dummies (with 2004 regarded as the base year). The joint significance test of the time effects is performed after 58 University of Ghana http://ugspace.ug.edu.gh estimating the specified model in equation (4.2) with the year dummies using OLS. The results of the OLS with year dummies and the time effect test are presented in Appendix (1.1) and (1.2) respectively. From Appendix (1.2), the result of the time effects test indicates that the year dummies are jointly statistically significant. The result from the test means that the pooled regression is not the appropriate model than the model with time effects. 5.2.2 Heterogeneity Bias One basic issue confronting the estimation of panel dataset through a pooled regression technique is heterogeneity bias. Hence, testing for the existence of individual heterogeneity of the trading partners is essential. The estimates in Table 5-2 below present the heterogeneity bias test through the adoption of Breusch and Pagan Lagrangian Multiplier test for random effects. The heterogeneity test serves as the basis for choosing among the pooled regression model and the random effect model. The result of the test in Table 5-2 reveals that the null hypothesis is rejected; hence, the random effect model is preferred over the pooled regression. Table 5- 2: Breusch and Pagan Lagrangian multiplier test for random effects Variance sd = sqrt(Var) InEXPORgj 7.593999 2.755721 E 1.824347 1.350684 U 2.300172 1.516632 Test: Var(u) = 0 chibar2(01) = 1290.63 Prob > chibar2 = 0.0000 NB: * p < 0.10, ** p < 0.05, *** p < 0.01 Source: Author’s Computation using STATA. 5.2.3 Heteroscedasticity For the estimated results of the study to be reliable, the error terms must have equal and constant variance. To test for the heteroscedasticity, we used the Breusch - Pagan test to test the null 59 University of Ghana http://ugspace.ug.edu.gh hypothesis of constant variance. The results of the test from Table 5-3 indicate that the null hypothesis of constant variance is rejected and affirms the existence of heteroscedasticity. Table 5- 3: Breusch-Pagan / Cook-Weisberg test for heteroskedasticity Ho: Constant variance Variables: fitted values of InEXPORTij chi2(1) = 150.95 Prob > chi2 = 0.0000 NB: * p < 0.10, ** p < 0.05, *** p < 0.01 Source: Author’s Computation using STATA. 5.2.4: Park Test As noted in the previous chapter, the Park test permits one to test the hypothesis that constant elasticity of the model can be evaluated in the log-linear model. The results of the Park auxiliary regression and test are presented in Appendix (1.3) and Table 5-4 respectively. The test in Table 5-4 shows that the log-linear specification of the model can't be rejected; however, Silva and Tenreyro (2006) pointed out that the log-linearization of the equation is valid under very restrictive conditions on the conditional appropriation of the dependent variable. Hence, Park test is not sufficient in concluding that the log-linear model is more appropriate than the non-linear (PPML) model. Table 5- 4: Park test Test =0 F( 1, 231) = 0.45 Prob > F = 0.5011 N B: * p < 0.10, ** p < 0.05, *** p < 0.01 Source: Author’s Computation using STATA. 5.2.5: Gauss-Newton Regression (GNR) tests From chapter four of this study, we indicated that the Gauss-Newton Regression (GNR) test the hypothesis that the coefficient on isn't statistically significant, and it suggests that 60 University of Ghana http://ugspace.ug.edu.gh failing to reject the null hypothesis indicate that the PPML postulation that V = E can't be rejected, thus, justifying the utilization of PPML estimator over log-linear estimator. The results of the Gauss-Newton Regression and the GNR test are presented in Appendix (1.4) and Table 5-5 respectively. Table 5- 5: Gauss-Newton regression (GNR) tests Test =0 F( 1, 998) = 0.53 Prob > F =0.4673 N B: * p < 0.10, ** p < 0.05, *** p < 0.01 Source: Author’s Computation using STATA. From Table 5-5 above, the GNR test reveals that PPML assumption V = E cannot be rejected and this justified the use of PPML estimator. 5.2.7: Regression Equation Specification Error Test (RESET) As noted in the previous chapter, the RESET test enables us to verify whether the functional form of the estimated gravity equation is defined correctly. The RESET tests the null hypothesis of no misspecification and we utilized it to determine which of the estimation technique fit our data well. The results for the RESET test on both the log-linear and PPML model are presented in Table 5 -6 and Table 5 -7 below. Table 5- 6: RESET test for Log-Linear Model InEXPORT Coefficient _hat 0. 9471*** _hatsq -0.0263 ** NB: * p < 0.10, ** p < 0.05, *** p < 0.01 Source: Author’s Computation using STATA. Table 5- 7: RESET test for PPML model EXPORT Coefficient _hat 1.0370*** _hatsq -0.0163 NB: * p < 0.10, ** p < 0.05, *** p < 0.01 Source: Author’s Computation using STATA. 61 University of Ghana http://ugspace.ug.edu.gh The estimated results in Table 5-6 show that the null hypothesis of no misspecification was rejected under the log-linear model. This means that estimating the regression equation through logarithmic specification is not valid. On the other hand, we obtained the opposite result for the RESET test under the PPML, hence, the results in Table 5-7 indicate that the null hypothesis of no misspecification is not rejected. Given the presence of heteroscedasticity and zero trade values coupled with strong confirmation that the PPML estimator does not suffer from strong misspecification errors in view of RESET test, we infer that the estimated results from Poisson Pseudo Maximum Likelihood Estimator (PPML) model are reliable. 5.3: Discussion of the Empirical Results The estimates shown in Table 5-8 below of this study are obtained from estimating the specified model of equation (4.2) with the pooled regression, random effects model and fixed effects model, whereas the PPML estimates were attained from equation (4.6). Appendix (1.5) presents the Hausman test, it shows that the null hypothesis of no correlation between the regressors and individual heterogeneity is rejected. Hence, it indicates that estimating the specified model using the random effects over the fixed effects estimation technique is inappropriate. Although, the test on the Hausman implies that the fixed effect estimator is more appropriate than the random effect estimator, however, the RESET test shows that the log-linear estimation technique is inappropriate. The result of the RESET test confirms that of Covaci and Moldovan (2015). Since Gauss-Newton regression (GNR) test and the RESET test render the results from the log-linear 62 University of Ghana http://ugspace.ug.edu.gh estimates inappropriate, the study concentrates on the parameter estimates obtained under the Poisson Pseudo Maximum Likelihood Estimator (PPML) estimators. From Table 5-8 below, the performance of the model of interest was fairly satisfactory, since, the vast majority of the key gravity covariates had significant coefficients with the anticipated signs. The coefficient of determination is approximately 60%; although not fairly high, it indicates that the regressors satisfactorily explain 60% of the variations in NTEs in Ghana. In sum, the estimates of the model reveals that exporter and importer GDP, the bilateral distance, importer population, exporter remoteness, common language, common border, regional trading blocs, importer trade openness and importer institutional quality are all significant with the expected signs whereas exporter population, real bilateral exchange rate and the land area of the trading are not significant in determining NTEs in Ghana. 63 University of Ghana http://ugspace.ug.edu.gh Table 5- 8: Estimates of the gravity model using OLS, RE, FE, and PPML from 2004 – 2016 Independent variable Pooled OLS RE Model FE Model PPML Model Log (GDPgt) 3.945 4.924 5.168 2.727 (-1.991)** (-1.457)*** (-1.481)*** (-1.294)** L og (GDPjt) 0.726 0.875 2.23 0.492 (-0.0865)*** (-0.183)*** (-0.567)*** (-0.0701)*** L og (POPgt) -7.639 -10.64 -14.49 -4.5 (-5.769) (-4.174)** (-4.229)*** (-3.728) L og (POPjt) 0.495 0.345 2.524 0.439 (-0.0996)*** (-0.246) (-1.085)** (-0.0737)*** L og (DISTANCEgj) -1.136 -1.196 -0.821 (-0.128)*** (-0.343)*** (-0.0851)*** L og (AREAj) -0.121 -0.266 -0.0752 (-0.0497)** (-0.14)* (-0.031)** Log (RLjgt) 0.193 -0.0573 - 0.0267 0.795 (-0.164) (-0.134) (-0.14) (-0.161)*** L og (POLSTAjgt) 0.177 0.00656 -0.0718 0.131 (-0.0919)* (-0.141) (-0.156) (-0.0795)* Log (RQjgt) -0.151 0.27 0.346 -0.431 (-0.24) (-0.27) (-0.328) (-0.12)*** L og (RBERgjt) 0.0354 0.00782 -0.00455 0.00853 (-0.0358) (-0.0918) (-0.293) (-0.0433) COMLANgj 0.518 0.543 0.429 (-0.142)*** (-0.418) (-0.0964)*** LANDLOCKgj -0.8 -0.804 0.812 (-0.287)*** (-0.629) (-0.135)*** E Ugj 0.67 0.463 1.008 (-0.204)*** (-0.553) (-0.115)*** ECOWASgj 3.603 3.583 1.882 (-0.235)*** (-0.742)*** (-0.204)*** Log (OPENNESSjgt) 0.685 -0.0652 -0.257 0.806 (-0.165)*** (-0.276) (-0.336) (-0.131)*** BORDgj 0.504 0.444 0.956 (-0.301)* (-1.151) (-0.126)*** C onstant Included Yes Yes Y es Yes No. of Observation 983 983 983 1001 No. of Importing countries 78 78 78 78 NB: Standard errors are in parentheses and ‘***’ means p<0.01, ‘**’ means p<0.05 and ‘*’ means p<0.10. Also, g and j refer to Ghana and importing countries respectively. Source: Authors own computation with the help of STATA. 64 University of Ghana http://ugspace.ug.edu.gh 5.3.1 Economic Size Gross Domestic Product (GDP) From column 4 of table 5-8, estimates on GDP reveals that the supply of NTEs from Ghana increases whenever there is an expansion in domestic and the trading partner GDP. The outcome on the importer’s GDP indicates that NTEs is a normal good and demand for it is fairly income elastic. On the other hand, the estimate on domestic GDP indicates that the Ghanaian economy becomes more productive as income increases. The outcome on GDP conform with the theoretical results of Chan et al., (2007), ), Jordaan & Eita (2011), Turkson (2012), Karamuriro & Karukuza (2015), Braha et al., (2016), Dlamini et al., (2016) and Kwame (2017) that GDP positively affect the export growth of a country. Population The estimate on population from column 4 of table 5-8 shows that importer's population has a positive and significant effect on Ghana's NTEs. However, the estimate on Ghana’s population reveals that it has a negative effect on the supply of NTEs, but is statistically insignificant. The outcome on population support the findings of Braha et al. (2016), Dlamini (2016), Karamuriro & Karuza (2015), Jordaan & Eita (2011), Hatab et al., (2010) and Chan & Au (2007). 5.3.2 Distance The estimate on the distance variable indicates that distance has a significant negative effect on Ghana’s NTEs. This outcome conforms to theoretical conclusion that transport cost adversely influences trade. The outcome on distance supports the results of Kwame (2017), Braha et al. 65 University of Ghana http://ugspace.ug.edu.gh (2016), Dlamini (2016), Karamuriro & Karuza (2015), Turkson (2012) and Hatab et al., (2010), but contradicts the findings of Jordaan & Eita (2011). 5.3.3 Control Variables The estimate reveals that dummy variables such as common border, common official language and the landlocked nature of the trading partner significantly influence Ghana’s NTEs. Common border and common official language exert positive effects on NTEs from Ghana. The results from column 4 in table 5-8 show that exporting non-traditional products to countries it has common border and language enhance exports significantly by 160 percent and 54 percent respectively as compared to countries it shares no border and official language. These results are consistent with the findings of Eita & Jordaan (2011), Turkson (2012) and Karamuriro & Karukuza (2015) whereas the estimates on common border and official language contradict the findings of Hatab et al., (2010) and Braha et al., (2016) respectively. As pointed out by Melitz (2007), the existence of linguistic barriers among trading partners can be a key obstacle to bilateral trade. The landlocked nature of importing countries exerts positive effects on NTEs from Ghana. The outcome on the landlocked nature of the trading partner contradict the findings of Anderson & Van Wincoop (2003), Jansen & Piermartini (2009) and Braha et al. (2016) that transport costs are higher with a landlocked trading partner. The reason behind this contradiction could be attributed to the fact that the top importer of Ghana’s non-traditional products in Africa currently is a landlocked country, thus Burkina Faso (GEPA, 2016), and its shares a common border with Ghana; hence, the cost associated with exporting non-traditional products is low. 66 University of Ghana http://ugspace.ug.edu.gh 5.3.4 Regional Trading Bloc Variables The estimate on European Union (EU) variable reveals that Ghana’s NTEs increase significantly when there is an expansion of trade integration between Ghana and member states of the Union. The finding on European Union trade bloc supports the theoretical conclusion that regional trade expansion facilitates trade flows between countries. This variable facilitates Ghana’s NTEs through the provision of market opportunities and trade policy such as the Economic Partnership Agreements (EPAs) which ease the cost associated with trade. The estimate confirms the findings of Thursby and Thursby (1987), Chan et al., (2007), Jordaan and Eita (2011), Habab et al., (2010) and Braha et al., (2016). The result on the Economic Community of West African States (ECOWAS) indicates that Ghana’s NTEs increase significantly when there is an expansion of trade relations between Ghana and member states in the community. This outcome on ECOWAS affirms theoretical conclusion that regional trading blocs facilitate trade flows between countries. ECOWAS promotes Ghana’s NTEs through the provision of market opportunities and trade policy such as the ECOWAS Trade Liberation Scheme (ETLS) that promotes free trade within the member states, hence, easing the cost of exports. The outcome of ECOWAS supports the findings of Turkson (2012). The estimate on the Association of Southeast Asian Nations (ASEAN) variable shows that Ghana’s NTEs improves very significantly whenever there is an expansion of trade integration between Ghana and member countries from ASEAN. The finding on ASEAN supports theoretical conclusions that the expansion of trade with regional trade union improves trade flows. 67 University of Ghana http://ugspace.ug.edu.gh 5.3.4 Institutional Quality The basic gravity models of trade often predict that the institutional quality of both the importer and the exporter affects international trade positively (Linders et al., 2005). Hence, institutional environment is commonly defined as a significant factor in reducing the level of uncertainty associated with trade between nations (Jansen and Nordas, 2004). The results of our estimates confirm the expected impact of institutional variables such as regulatory quality (RQ), rule of law (RL) and political stability (POLSTA). The estimates on the importer’s rule of law (RL) and political stability (POLSTA) from column 4 of table 5-8 shows that Ghana’s NTEs improves very significantly whenever there is an expansion in such variables whereas the estimate on importer’s regulatory control (RQ) significantly influence NTEs negatively. The results support the theoretical conclusion from the gravity model literature that the advancement in institutional quality tends to significantly influence trade flows. The negative effect of the trading partner institutional variable such as regulatory quality control could be attributed to the poor performance of Ghanaian institutions such as Ghana Standards Authority (GSA), Ghana Exports Promotion Council (GEPC) and the Food and Drug Authority (FDA) in ensuring that products exported from the country are safe, reliable and are of good quality. Hence, when trading partners tighten institutions responsible for products quality checks, it affects NTEs from Ghana negatively. These outcomes are in conformity with the work of Braha et al (2016). The estimate on the importer’s trade openness shows that NTEs increase significantly whenever there is an expansion of trade with a more trade open country. The outcome of trade openness support the findings of Boansi et al., (2014), but contradict that of Braha (2016) and Dlamini (2016). 68 University of Ghana http://ugspace.ug.edu.gh 5.3 Prediction of Ghana’s Non-Traditional Exports Potential The estimates on the overall mean Actual, Potential and Unexhausted for Ghana’s NTEs over the study period are presented in Table 5-9 below. The results were obtained through the estimation of equation (4.12). Table 5- 9: Overall Ghana's mean Actual and Potential NTEs for the entire period under study (2004 – 2016). Actual NTEs Potential NTEs Unexhausted NTEs The speed of (US Dollar ) (US Dollar ) (US Dollar ) Convergence/Divergence (%) 44244630363 44155577581 89052782 -0.2013 Source: Author’s calculations The outcome in Table 5-9 reveals that Ghana’s mean of potential Non- Traditional Exports (NTEs) was worth over US$ 44.1 billion compared to over US$ 44.2 billion worth of mean of actual Non-Traditional Exports during entire period under study. The un-exhausted exports potential for NTEs over the 13 years was estimated to worth about USS$ 89.0 million. Also, the estimate reveals that Ghana is having exports convergence with the world since average actual export is going faster than that of the potential export. The estimates of the speed of convergence of Ghana and her individual trading partners are presenting Appendix (1.6). The result in Table 5-9 is an indication that, there is an urgent need for the restructuring of socio-economic and institutional factors that are vital to exports diversification in order to expand the NTEs market. However, when the country-specific exports potentials were simulated, the results uncovered that Ghana has NTEs potentials with 45 out of the 78 countries under study. Figure 5.1 shows the overall mean of actual, potential and unexhausted NTEs of the top 10 trading partners over the study period. Figure 5.1 below shows that Germany had the highest unexhausted exports worth 69 University of Ghana http://ugspace.ug.edu.gh over USS$ 0.921 billion followed by Ireland (USS$ 0.439 billion), Italy (USS$ 0.347 billion), Luxembourg (USS$ 0.286 billion), Nigeria (USS$ 0.245 billion), Cote D’ Voire (USS$ 0.236 billion), Portugal (USS$ 0.176 billion), Singapore (USS$ 0.155 billion), Poland (USS$ 151 billion and Hong Kong (USS$ 0.150 billion). Also, Figure 5.2 below shows the top ten countries with the most exhausted trade potential. From Figure 5.2, the country with the highest exhausted NTEs is Netherland followed by France, United Kingdom, India, Vietnam, Burkina Faso, Estonia, United States, Malaysia and Switzerland. Figure 5. 1: Overall Top 10 Unexhausted Exports Destination for Ghana NTEs HONG KONG POLAND Unexhausted NTEs SINGAPORE Potential NTEs PORTUGAL Actual NTEs COTE D'IVOIRE NIGERIA LUXEMBOURG ITALY IRELAND GERMANY 0.000 0.500 1.000 1.500 2.000 US$ Billion Source: Author’s calculations Figure 5. 2: Overall Top 10 Exhausted Exports Destination for Ghana NTEs 70 Countries University of Ghana http://ugspace.ug.edu.gh SWITZERLAND MALAYSIA UNITED STATES ESTONIA BURKINA FASO Exhausted NTEs VIETNAM Potential NTEs INDIA Actual NTEs UNITED KINGDOM FRANCE NETHERLANDS -2.000 -1.000 0.000 1.000 2.000 3.000 Billion US Dollars Source: Author’s calculations The goal of every economy is to achieve its full trade potentials through trade engagement process or even through unilateral reforms. The outcome of the trade potential above will serve as a basis in the engagement of bilateral and multilateral trade decisions of the economy, hence minimizing the effect of existing trade restrictive measures that are hindering the growth of NTEs. The estimates show that it is important to implement trade policies that will promote exports to countries with unexhausted exports values to enable the exploitation of exports potential. Also, the estimate on the exhausted export's destination is an indication that exporting NTEs to those countries is associated with high cost to the economy. But, as noted by Jordaan and Eita (2011), further analysis of each country is needed to determine and identify possible factors that are responsible for such outcomes. 5.4 Conclusion The results of the study uncovered that 15 out of 17 variables used to explain the growth rate of Non-Traditional Exports from Ghana. Furthermore, the estimates on 14 out of the 15 variables that explain NTEs growth affirms their respective theoretical conclusions, however, the estimate 71 countries University of Ghana http://ugspace.ug.edu.gh on the landlocked nature of the trading partner contradict its theoretical conclusions. Finally, the estimates on NTEs potentials shows that Ghana’s NTEs have exports potentials with 45 out of the 78 countries used in the study. 72 University of Ghana http://ugspace.ug.edu.gh CHAPTER SIX CONCLUSION AND RECOMMENDATION 6.1 Introduction The focal point of this chapter is to discuss the summary and conclusions of the study and further make available policy measures that are based on the findings of the study. 6. 2 Summary of Findings The objective of this study was to examine factors determining Ghana's NTEs growth by applying the gravity model of trade approach. The aim of this research is to acquire knowledge of export diversification and Non-Traditional Exports growth to provide guidelines for future export opportunities. The study used a yearly panel data on Ghana and her 78 major trading partners from 2004-2016 with Ghana’s NTEs value as the dependent variable and 17 independent variables. These 78 countries were selected based on their magnitude of trade with Ghana and the accessibility of the required panel data on them. Apart from the basic variables of the gravity model of trade, additional variables such as real bilateral exchange rate, importer openness to trade, importer institutional quality, land area of the trading partner, dummy variables such as common border, common language, landlocked nature of the trading partner and regional trading blocs dummies like EU, ECOWAS and ASEAN were included to improve the consistency and the integrity of the model in explaining the basic characteristics of factors influencing Non-Traditional Exports growth in Ghana. 73 University of Ghana http://ugspace.ug.edu.gh The regression was performed in four-panel data approaches; the Pooled regression, Random Effect Model (REM), Fixed Effect Model (FEM) and Poisson Pseudo Maximum Likelihood estimator (PPML). However, the robustness checks such as Gauss-Newton regression test and the RESET test suggested that the PPML estimator is the most appropriate model compared to the log-linearized models such as the Pooled regression, Random Effect Model, and Fixed Effect Model. The results from the PPML shows that Ghana’s GDP, Importer’s GDP, Importer’s population, institutional variables like political stability and the rule of law of the importer and the importer’s trade openness have a positive significant impact on the growth rate of Non- Traditional Exports from Ghana. However, variables such as bilateral distance, land area sizes of the importer and institutional variable like the regulatory quality of the importer exert negative effects on the NTEs growth from Ghana. Furthermore, the study reveals that regional trading blocs dummy variables such as EU, ECOWAS, ASEAN and other dummy variables like landlocked nature of the trading partner, sharing a common border and official language (English) significantly facilitates Ghana’s NTEs growth. Finally, variables such as the bilateral exchange rate and the domestic population were statistically insignificant in explaining the growth of Ghana’s NTEs. Finally, the study reveals that Ghana’s NTEs have unexploited potentials with 45 out of the 78 trading partners used with Germany, Ireland, Luxembourg, Nigeria, Cote D’Ivoire, Portugal, Singapore, Poland, Hong Kong and Greece being the top 10 countries. On the other hand, countries such Netherland, France, United Kingdom, India, Vietnam, Burkina Faso, Estonia, United States, Malaysia and Switzerland are the topmost 10 exhausted NTEs destinations. The rest of the 35 NTEs potential destinations and 23 exhausted NTEs potentials are presented in Appendix (1.6). 74 University of Ghana http://ugspace.ug.edu.gh 6.3 Recommendations Based on the estimates obtained from the estimation of the specified model in equation (4.6) of the study, the following recommendations are suggested:  The estimates of the study show that Ghana’s Non-Traditional Exports are positively influenced by the expansion of Ghana’s GDP, importer’s GDP and population. Therefore, policies that lead to a remarkable growth of Ghana and her trading partner’s economy should be enhanced so that GDP of Ghana and the importers will increase. Policies such as tax cut, interest rate cutting, privatization, deregulation and the expansion of government expenditure through the provision of improved education and infrastructure could help expand GDP growth, hence, propel NTEs growth in Ghana. Since the estimated result from trading partner’s population indicates that the trading partner’s population is a net consumer of NTEs, policymakers in Ghana should implement policies that will expand Non-Traditional Exports with populated countries such as China, India, United States, Indonesia, Brazil, Pakistan, Nigeria and others to exploit market opportunities associated with larger population countries.  The formations of regional trading blocs such as ECOWAS, EU, and ASEAN trade blocs have a significant positive effect on Ghana’s Non-Traditional Exports. Hence, the study recommends that policymakers in Ghana should initiate trade policies that will improve and expand trade integration with these regional trading blocs to provide the needed markets for Ghana’s non-traditional products.  The importer’s institutional quality such as rule of law and political stability affect Ghana Non-Traditional Exports positively. Hence, the study recommends that exports from 75 University of Ghana http://ugspace.ug.edu.gh Ghana should be targeted to countries with high level of rule of law and political stability. However, the study revealed that institutional quality variables such as the regulatory quality of the importer had a negative impact on NTEs from Ghana. As noted in Chapter Five of this study, the negative effect of the trading partner’s regulatory quality control could be credited to poor performance of domestic institutions such as Ghana Standard Authority (GSA), Food and Drugs Authority (FDA) and others responsible for the exportation of safe, reliable and good products from the country. The performance of these institutions could be attributed to inadequate support base from government and other organizations in Ghana. Hence, the study recommends that trade policies of the Ghanaian Government should regard the improvement of such institutions a focal objective. This will enable such institutions to implement better oriented corrective measures to improve upon the inspection and control system for exportable products at both production and exit points to avoid restriction and ban of Ghanaian products by the importers.  The results from the study show that geographical factor such distance influence Ghana’s NTEs negatively. As the distance between Ghana and the destination country gets larger, so does the transportation costs. Hence, the study recommends that other things being equal, Ghana’s Non-Traditional products should be exported to closer countries in order to reduce the cost of transportation and increase the valued exported.  The results from NTEs potential indicate that policymakers should implement initiatives that will enhance Ghana’s NTEs to countries with unexhausted trade potentials. These could be made possible through the enhancement of trade relationships with those countries. Also, to enhance trade with countries that have exhausted trade, policymakers 76 University of Ghana http://ugspace.ug.edu.gh should initiate programs that will lead to NTEs diversification to enable recapturing exhausted market destination. 6.4 Limitation of Study The study attempted to ascertain determinants of Ghana’s Non-Traditional Exports to 78 major trade partners from 2004 to 2016, data employed was not disaggregated at the sector level. The data was aggregated as overall Non-Traditional Exports to the trading partners. However, NTEs can be decomposed on sectors basis as mention in previous chapters of this study. We recommend that future studies in this area should disaggregated data at sector level to distinguish and identify the specific determinants of Non-Traditional Exports at sector level Furthermore, due to difficulties associated with getting data on prices, the study used the remoteness index as a proxy for the multilateral resistance in trade. 77 University of Ghana http://ugspace.ug.edu.gh References Addo, E., & Marshall, R. (1998). Ghana non-traditional export sector: expectations,achievements and policy issues. Geoforum 31 (2000) 355-370,www.elsevier.com/locate/geoforu. Agyei-Sasu, F., & Egyir, I. S. (2010). Tobit Estimation of the Intensity of Export Success of Horticultural Enterprises in Ghana. University Of Ghana, Legon, Accra. Ampadu-Agyei, O. (1994). Non-Traditional Agricultural Exports and Natural Resource Management in Ghana:Practices and Prospects. Center for International Development and Environment.World Resources Institute,1709 New York Avenue, NW, Suite 700 Washington, DC 20006, 2-4. Anagaw, K. B., & Demissie, M. W. (2013). Determinants of Export Performance in Ethiopia: Var Model Analysis. National Monthly Refereed Journal of Research in Commerce & Management, Anderson, J. (1979). A Theoretical Foundation for the Gravity Equation. The American Economic Review, Vol. 69, 106-116. Anderson, J. E. (2004). International Trade Theory. Boston College. Boston College., 2-6. Anderson, J. E., & Wincoop, E. V. (2003). Gravity with Gravitas: A Solution to the Border Puzzle. The American Economic Review, Vol. 93, issue 1, Nashville, 170-192. Bacchetta, M., Beverelll, C., Cadot, O., Fugazza, M., Grether, J. M., Helble, M., . . . Plemartini, R. (2012). A Practical Guide to Trade Policy Analysis. New York, United States of America, United Nations Publications. Balassa, B. (1963). An Empirical Demonstration of Classical Comparative Cost Theory. The Review of Economics and Statistics, Vol. 45, No. 3, 231-238. 78 University of Ghana http://ugspace.ug.edu.gh Bergstrand, H. J. (1985). The Gravity Equation in International Trade: Some Microeconomic Foundations and Empirical Evidence. Review of Economics and Statistics, Vol. 67, No. 3, 474–481. Bergstrand, H. J. (1989). The Generalized Gravity, Monopolistic Competition, and the Factor- Proportions Theory in International Trade. The Review of Economics and Statistics, MIT Press. Vol. 71, No. 1, 143-153. Bergstrand, J., Egger, P., & Larch, M. (2013). "Gravity Redux: Estimation of gravity-equation coefficients, elasticities of substitution, and general equilibrium comparative statics under asymmetric bilateral trade costs". Journal of International Economics, Volume 89, Issue 1, pp 110-121. Boansi, D., Okonon, B. O., & Appah, J. (2014). Determinants of Agricultural Export Trade: Case of Fresh Pineapple Exports from Ghana. British Journal of Economics, Management & Trade, 4(11): 1736-1754, 2014. Braha, K., Qineti, A., Lazorcakova, E., & Galik, J. (2016). Determinants of Albanian Agricultural Export: The Gravity Model Approach. Dep Working Paper Series, 6-15. Brulhart, M., & Kelly, M. J. (1999). Ireland's trading potential with central and eastern European countries: A gravity study. Economic and Social Review, 30(2), 159 – 174. Burger , M. F., & Linders, G. J. (2009). On the Specification of the Gravity Model of Trade: Zeros, Excess Zeros and Zero-Inflated Estimation. Spatial Economic Analysis 4(2), 167- 90. Carbaugh, R. J. (2009). International Economics, 12th Edition. OH,USA: South-Western Cengage Learning. 79 University of Ghana http://ugspace.ug.edu.gh Chan A, E. M., & Au a, K. F. (2007). Determinants of China's textile exports: An analysis by gravity model. The Journal of The Textile Institute. Charles, H. W. (2009). International Trade Theory PPT Chapter 5. Global Business Today, The McGraw-Hill Companies, Inc., 1-65. Christie, E. (2002). Potential Trade in Southeast Europe: A Gravity Model Approach. Working Paper, The Vienna Institute for International Economic Studies-WIIW downloaded on 14 March 2002. Covaci, G., & Moldovan, S. (2015). Determinants of service exports of Lithuania: a gravity model approach. SSE Riga Student Research Papers 2015:1(166). Darko-Opoku, A. O. (2014). Ghana Free Zones Programme: Providing a Conducive Investment Environment. GFZP, 5-9. Deardorff, A. (1985). Determinants of Bilateral Trade: Does Gravity Work in a Classical World? Paper Presented at Conference on The Regionalization of the World Economy NBER, Woodstock, Vermont. Dlamini , S. G., Abdi-Khalil, E., Alexander , R., & Micah , B. (2016). Determinants of Swaziland’s Sugar Export: A Gravity Model Approach. Online Published. Dunn, R. M., & Mutti, J. H. (2004). International Economics, Sixth edition. New York: Routledge. Egger, P. (2000). A Note on the Proper Econometric Specification of the Gravity Equation. WIFO Working Papers 108. Egger, P. (2002). An Econometric View on the Estimation of Gravity Models and the Calculation of Trade Potentials. 80 University of Ghana http://ugspace.ug.edu.gh Egger, P., & Pfaffermayr, M. (2004). Distance, Trade and FDI: A SUR Hausman-Taylor Approach. Journal of Applied Econometrics 19 (2), 227-246. Eichengreen, B., & Irwin, D. A. (1998 ). "The Role of History in Bilateral Trade Flows", in Jeffrey A. Frankel, ed., The Regionalization of the World Economy. Chicago: University of Chicago Press, pp. 33-57. Eita, J. H. (2008). Determinants of Namibian Exports: A Gravity Model Approach. University of Namibia, Namibia. Fosu, A. K. (2001). Emerging Africa: The Case of Ghana. African Economic Research Consortium, 1-4. Frankel, J. A. (1997). Regional Trading Blocs in the World Economic System. Chicago, IL: University of Chicago Press. Glick, R., & Rose, A. (2002). Does a currency union affect trade? The time series evidence. European Economic Review, 46(6), 1125-1151. Gomez , E. H. (2013). Comparing alternative methods to estimate gravity models of bilateral trade. The University of Granada, Campus de la Cartuja s/n, 18071 Granada, Spain. Gomez, E. H., & Milgram, J. B. (2010). Are estimation techniques neutral to estimate gravity equations? An application to the impact of EMU on third countries’ exports. Universidad de Granada,Facultad de Ciencias Económicas y Empresariales. Classification JEL Code. C13, C33, F10, F15, 1-29. Greene, W. H. (2008). Econometric Analysis. Sixth Edition. New Jersey ,07458: Pearson Prentice Hall, Upper Saddle River. Habab, A. A., Romstad, E., & Huo, X. (2010). Determinants of Egyptian Agricultural Exports: A Gravity Model Approach. Modern Economy, 134-143. 81 University of Ghana http://ugspace.ug.edu.gh Hamilton, C. B., & Winters, A. L. (1992). Opening Up International Trade with Eastern Europe. Economic Policy, 14, 77 – 116. Helpman, E., & Krugman, P. (1985). Market Structure and Foreign Trade, Increasing Returns, Imperfect Competition, and the International Economy. MIT Press, Cambridge MA/London. Helpman, E., Melitz, M., & Rubinstein, Y. (2008). Estimating trade flows: trading partners and trading volumes. Quarterly Journal of Economics (73), 441–486. ISSER. (2013). The State of the Ghanaian Economy. The University of Ghana. Jakab, Z. M., Mihaly, K., & Oszlay, A. (2001). How Far Has Trade Integration Advanced?: An Analysis of the Actual and Potential Trade of Three Central and Eastern European Countries. Journal of Comparative Economics 29(2):, 276-292. Jansen, M., & Nordas, H. K. (2004). Institutions, trade policy and trade flows. WTO Staff Working Paper, No. ERSD-2004-02. Jordaan, A. C., & Eita, J. H. (2011). Identifying South Africa's Wood Exports Potential Using a Gravity Model. 2010 International Conference on E-business, Management and Economics, IPEDR vol.3 (2011) © (2011) IACSIT Press, Hong Kong. Karamuriro, W. H., & Karukuza, N. (2015). Determinants of Uganda’s Export Performance: A Gravity Model Analysis. International Journal of Business and Economics Research, 50- 52. Kaur, S., & Nanda, P. (2011). An Analysis of Actual and Potential Exports of Pakistan with Saarc Countries: A Panel Data Analysis. Pakistan Journal of Applied Economics, Vol. 21 Nos. 1 & 2, 69-91. 82 University of Ghana http://ugspace.ug.edu.gh Kepaptsoglou, K., Karlaftis, M. G., & Tsamboulas, D. (2010). The Gravity Model Specification for Modelling International Trade Flows and Free Trade Agreement Effects: A 10-year Review of Empirical Studies. The Open Economics Journal, 3, 1-13. Kristjánsdóttir, H. (2005). A Gravity Model for Exports from Iceland. Centre for Applied Microeconometrics, Department of Economics -University of Copenhagen, 9. Kyereboah-Coleman, A., & Biekpe, N. (2006). The link between corporate governance and performance of the non-traditional export sector: evidence from Ghana. Research paper. Linders, G. J., Slangen, A., De Groot, H. L., & Beugelsdijk, S. (2005). Cultural and institutional determinants of bilateral trade flows. Tinbergen Institute Discussion Paper. Linneman, H. (1966). An Econometric Study of International Trade Flows. North Holland Amsterdam. Liu, X. (2009). GATT/WTO promotes trade strongly: Sample selection and model specification. Rev Int Econ 17, 428-446. Lwesya, F. (2016). Export Diversification Dynamics in Tanzania Non-Traditional Products Approach. Journal of Economics and Sustainable Development. ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online), Vol.7, No.13, 24. MacDougall, G. D. (1951). British and American Exports: A Study Suggested by the Theory of Comparative Costs. Part I. The Economic Journal, Vol. 61, No. 244 (Dec. 1951), 697- 724. Makochekanwa, A., & Jordaan, C. A. (2008). Identifying the trade theory model behind Botswana’s sectoral exports. 13th Annual Conference on Economic Modeling in Africa, 5-6. 83 University of Ghana http://ugspace.ug.edu.gh Manning, W. G., & Mulahy, J. (2001). Estimating Log Models: to Transform or not to Transform? Journal of Health Economics , 20: 461-494. Martin , W., & Pham, C. (2008). Estimating the Gravity Equation when Zero Trade Flows are Frequent. MPRA Working Paper 9453, University Library of Munich. Martinez-Zarzoso, I. (2013). The log of gravity revisited. Applied Economics, 45:3, 311-327. Martinez-Zarzoso, I., Marquez-Ramos, L., & Saurez-Burguet, C. (2007). The role of distance in gravity regressions: is there really a missing globalization puzzle? The BE Journal of Economic Analysis and Policy, 7, 1–25. Matyas, L. (1998). The Gravity Model: Some Econometric Considerations. World Economy, 21: 397–401. doi:10.1111/1467-9701.00136. McCallum, J. (1995). National Borders Matter: Canada-U.S. Regional Trade Patterns. American Economic Association, The American Economic Review, Vol. 85, No. 3. http://www.jstor.org/stable/2118191, pp. 615-623. Medina-Smith, J. E. (2008). Is the Export-Led Growth Hypothesis Valid for Developing Countries ?A Case Study of Costa Rica. United Nations Conference on Trade and Development (UNCTAD), Policy Issues in International Trade and Commodities. Melitz, M. (2007). North, South and Distance in the gravity equation. European Economic Review 51, 971-991. Moon, S., Sekwat, A., & Stanley, R. E. (2004). Testing the Assumptions of Pooled Time Series Cross-Sectional Regression Analysis through ARIMA and Interrupted Time Series Modelling. Beck and Katz Revisited. Paper prepared for the Western Social Science Conference. 84 University of Ghana http://ugspace.ug.edu.gh Nomfundo , P. V., & Nicholas , O. (2017). A Review of Imports Structure and Reforms in Ghana. E ur o E c o n om I ca, 146. Oduro, A., Asante, Y., Tsikata, G., & Jebuni, C. (1192). Diversifying Exports :The Supply Response of Non-Traditional Exports to Ghana's Economic Recovery Programme. London: Overseas Development Institute. Oguledo, V., & Macphee, C. (1994). Gravity Models: A Reformulation and an Application to Discriminatory Trade Arrangements. Applied Economics, 26, 107-120. Opoku, M. (2015). The Economic Partnership Agreement (EPA) Between West Africa and the European Union. MRBT, MoTI, 1-5. Osei-Assibey, E. (2015). Export Promotion in Ghana. University of Ghana, 5. Osei-Assibey, K. (2017). Exchange Rate Volatility, Earnings Uncertainty and Bidirectional Trade Flow: Empirical Evidence on Ghana. International Economic Journal, 12-18. Paas, T. (2000). Gravity Approach for Modeling Trade Flows between Estonia and the Main Trading Partners. Working Paper, No. 721, Tartu University Press, Tartu. Poyhonen, P. (1963). A Tentative Model for the Volume of Trade between Countries. Weltwirtschaftliches Archiv, vol. 90, pp. 93-99. Rahman, M. (2009). Australia’s Global Trade Potential: Evidence from the Gravity Model Analysis. Working Paper, Oxford Business & Economics Conference, Oxford., 1-40. Rahman, M. M. (2003). A Panel Data Analysis of Bangladesh’s Trade: The Gravity Model Approach. University of Sydney, NSW 2006, Australia, 2. Rose, A., & Wincoop, E. v. (2001). National Money as a Barrier to International Trade: The Real Case for Currency Union. American Economic Review, 91(2), 386-390. Salvatore, D. (1998). International Economics, 6th Edition, New Jersey, Prentice-Hall. 85 University of Ghana http://ugspace.ug.edu.gh Santos Silva , J. M., & Tenreyro, S. (2006). The Log of Gravity. Rev Econ Stat 88, 641-658. Santos Silva, J., & Tenreyro, S. (2011). Further simulation evidence on the performance of the Poisson pseudo-maximum likelihood estimator. Economics Letters, 112, 220–2., 220-2. Skosana, S. V., & Kabuya, F. I. (2014). An Empirical Analysis of Determinants of Swaziland’s Export Performance. International Journal of Sciences: Basic and Applied Research . Soren, P., & Bruemmer, P. (2012). Bimodality & the Performance of PPML. Institute forAgriceconomics Discussion paper 1202, Georg-August UniversitätGöttingen Germany. Stait, F. (2005). Are Exports the Engine of Economic Growth? An Application of Cointegration and Causality Analysis for Egypt, 1977-2003. Working Paper, African Development Bank, No. 76. Stern, H. (1962). The Significance of Impulse Buying Today . Journal of Marketing Vol. 26, No. 2 , 59-62. Stewart, J. (1948). Potential of Population and its Relationship to Marketing : Theory in Marketing. R. Cox and W. Alderson (Eds) ( Richard D. Irwin, Inc.,Homewood, Illinois). Suranovic, S. (2006). International Trade Theory and Policy Analysis Lecture Notes. Washington D. C.,The International Economics Study Center, © 1997-2006. http://internationalecon.com/v1.0. Thelle, M. H., Jeppesen, T., & Gjodesen-Lund, C. (2015). Export Performance and Poverty Reduction. Directorate-General for International Cooperation and Development, EU Development Policy and International Cooperation, Policy and Coherence, Economic Analysis Team, 11. 86 University of Ghana http://ugspace.ug.edu.gh Thrupp, L. A., Bergeron, G., & Waters, W. F. (1995). Bittersweet Harvests for Global Supermarkets: Challenges in Latin America’s Agricultural. Baltimore, Md.: World Resources Institute, 58-61. Thursby, G. J., & Thursby, C. M. (1987). Bilateral Trade Flows, The Linder Hypothesis, and Exchange Risk. Department of Economics, University of Michigan. Tinbergen, J. (1962). Shaping the World Economy: Suggestions for an International Economic Policy. New York, The Twentieth Century Fund. Turkson, F. E. (2012). Trade Agreements and Bilateral Trade in Sub-Saharan Africa (SSA): Estimating the ex-Post Trade Effects of the EU-ACP Preferential and Regional Trade Agreements. Work in Progress, 20-26. Wang, Z., & Winters, A. (1992). The Trading Potential of Eastern Europe. Journal of Economic Integration, 7: 113 - 136. Westerlund, J., & Wilhelmsson, F. (2011). Estimating the gravity model without gravity using panel data. Appl Econ 43, 641-649. Zhang, J., & Kristensen, G. (1995). A Gravity Model with Variable Coefficients: The EEC Trade with Third Countries’, Geographical Analysis. Vol. 27, 307-320. 87 University of Ghana http://ugspace.ug.edu.gh Appendix Appendix 1. 1: Time Effect OLS 88 University of Ghana http://ugspace.ug.edu.gh Appendix 1. 2: Time Effect Test ( 1) _IYEAR_2005 = 0 ( 2) _IYEAR_2006 = 0 ( 3) _IYEAR_2007 = 0 ( 4) _IYEAR_2008 = 0 ( 5) _IYEAR_2009 = 0 ( 6) _IYEAR_2010 = 0 ( 7) _IYEAR_2011 = 0 ( 8) _IYEAR_2012 = 0 ( 9) _IYEAR_2013 = 0 (10) _IYEAR_2014 = 0 (11) _IYEAR_2015 = 0 (12) _IYEAR_2016 = 0 F( 12, 956) = 3.24 Prob > F = 0.0001 Source: Author’s Computation using STATA. * p < 0.10, ** p < 0.05, *** p < 0.01 Appendix 1. 3: Park-type auxiliary regression and the Park test Coef. Std. Err. t P>t [90%Conf. Interval] 0.071559 0.10621 0.67 0.501 -0.137703 0.280822 Constant -0.92143 0.152289 -6.05 0.000 -1.221479 -0.62137 Test =0 F( 1, 231) = 0.45 Prob > F = 0.5011 S ource: Author’s Computation using STATA. * p < 0.10, ** p < 0.05, *** p < 0.01 Appendix 1. 4: Gauss-Newton regression (GNR) tests Coef. Std. Err. t P>t [90% Conf. Interval] 6.527364 2.966565 2.2 0.028 0.7059435 12.34878 -0.71196 0.979193 -0.73 0.467 -2.633472 1.209554 Constant -3.33976 2.637642 -1.27 0.206 -8.51572 1.836201 test =0 F( 1, 998) = 0.53 Prob > F = 0.4673 S ource: Author’s Computation using STATA. * p < 0.10, ** p < 0.05, *** p < 0.01 89 University of Ghana http://ugspace.ug.edu.gh Appendix 1. 5: Hausman Test . hausman fe Coefficients (b) (B) (b-B) sqrt(diag(V_b-V_B)) fe re Difference S.E. InGDPi 5.16753 4.92368 .24385 .2674939 InGDPj 2.230409 .8750803 1.355329 .5371479 InPOPi -14.48623 -10.63962 -3.846615 .6830477 InPOPj 2.524122 .3448658 2.179256 1.057146 InRLij -.0267179 -.0573087 .0305908 .0412337 InPOLSTAij -.0717817 .0065563 -.078338 .0658025 InRQij .3455823 .2698534 .075729 .185951 InRBERij -.0045483 .0078183 -.0123666 .2782941 InOPENNESSj -.256762 -.0652189 -.1915431 .1913568 b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg Test: Ho: difference in coefficients not systematic chi2(9) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 32.73 Prob>chi2 = 0.0001 (V_b-V_B is not positive definite) 90 University of Ghana http://ugspace.ug.edu.gh Appendix 1. 6: List of Countries used in the study with their overall actual, potential, unexhausted and speed of Convergence/divergence for NTEs over the period under study Srl Country Actual NTEs P o t e n tial NTEs U n e x h austed Speed of ( US $) (US $) NTEs (US $) Convergence/Divergence 1 GERMANY 742100570 1662906530 920805960 124.08 2 IRELAND 14984552 453641920 438657368 2927.40 3 ITALY 503886940 851341170 347454230 68.95 4 LUXEMBOURG 1621034 287638510 286017476 17644.14 5 NIGERIA 1407766640 1653012400 245245760 17.42 6 COTE D'IVOIRE 776917830 1012456730 235538900 30.32 7 PORTUGAL 74177595 249815000 175637405 236.78 8 SINGAPORE 132573426 287502133 154928707 116.86 9 POLAND 263203849 414446760 151242911 57.46 10 HONG KONG 23229627 174353276 151123649 650.56 11 GREECE 41721275 181774634 140053359 335.69 12 SWEDEN 21159493 146669528 125510035 593.16 13 THAILAND 20211223 138934076 118722853 587.41 14 FINLAND 15260558 131836267 116575709 763.90 15 DENMARK 233127117 343046850 109919733 47.15 16 CANADA 118714063 226098530 107384467 90.46 17 INDONESIA 18829780 126075899 107246119 569.56 18 RUSSIAN FEDERATION 24776128 117097964 92321836 372.62 19 KOREA 54082569 135757436 81674867 151.02 20 BRAZIL 88455730 160573140 72117410 81.53 21 NIGER 288168102 360028390 71860288 24.94 22 MEXICO 11282100 79943302 68661202 608.59 23 LITHUANIA 1751861 66009983 64258122 3667.99 24 CHINA 371989003 435474740 63485737 17.07 25 CROATIA 10176936 68470999 58294063 572.81 26 CHILE 8310149 62581464 54271315 653.07 27 SPAIN 802200550 855850450 53649900 6.69 28 REPUBLIC OF UGANDA 2544579 54246134 51701555 2031.83 29 ISRAEL 32569044 79656211 47087167 144.58 30 IRAN 15148347 54283139 39134792 258.34 31 TUNISIA 2794889 38603475 35808586 1281.22 32 CYPRUS 7023500 38751652 31728152 451.74 33 NORWAY 30749733 60063435 29313702 95.33 34 BOTSWANA 158631 26229729 26071098 16435.06 35 TOGO 1483596410 1501480120 17883710 1.21 36 JAPAN 163673132 181030728 17357596 10.61 37 TANZANIA 7400262 23452367 16052105 216.91 38 MOROCCO 27143597 42457560 15313963 56.42 39 PANAMA 5170366 19315630 14145264 273.58 91 University of Ghana http://ugspace.ug.edu.gh Appendix 1. 6 40 KENYA 7374883 18741736 11366853 154.13 41 ANGOLA 27035568 35806526 8770958 32.44 42 CHAD 6377692 10138484 3760792 58.97 43 SIERRA LEONE 65263945 68043012 2779067 4.26 44 NAMIBIA 7852389 9092679 1240290 15.80 45 PAKISTAN 20096129 20431761 335632 1.67 46 GUINEA BISSAU 7623063 5925244 -1697819 -22.27 47 PERU 16847306 14988058 -1859248 -11.04 48 MOZAMBIQUE 10090229 7196294 -2893935 -28.68 49 BELGIUM 570504240 560678820 -9825420 -1.72 50 GAMBIA 35713443 25194722 -10518721 -29.45 51 BAHAMAS 18978608 4027298 -14951310 -78.78 52 MAURITANIA 22574946 4043731 -18531215 -82.09 53 AUSTRALIA 88957209 69830426 -19126783 -21.50 54 LEBANON 29968961 9454517 -20514444 -68.45 55 CONGO 32836823 10924407 -21912416 -66.73 56 CAMEROON 64823775 41093679 -23730096 -36.61 57 EQUATORIAL GUINEA 45499739 21723972 -23775767 -52.25 58 GABON 45812601 15872240 -29940361 -65.35 59 GUINEA 75405428 35765664 -39639764 -52.57 60 LIBERIA 80088868 35634470 -44454398 -55.51 61 TURKEY 190670690 128319086 -62351604 -32.70 62 SOUTH AFRICA 158629905 90701483 -67928422 -42.82 63 UNITED ARAB EMIRATES 149697813 79834324 -69863489 -46.67 64 BENIN 426500310 355833030 -70667280 -16.57 65 MALI 434950259 360280180 -74670079 -17.17 66 EGYPT 146476886 71775187 -74701699 -51.00 67 SENEGAL 216000972 137763798 -78237174 -36.22 68 BULGARIA 180335554 81094343 -99241211 -55.03 69 SWITZERLAND 516159974 404579320 -111580654 -21.62 70 MALAYSIA 317890182 202219926 -115670256 -36.39 71 UNITED STATES 1021407160 844081170 -177325990 -17.36 72 ESTONIA 294262206 115850233 -178411973 -60.63 73 BURKINA FASO 1611134830 1357712310 -253422520 -15.73 74 VIETNAM 467570414 202342927 -265227487 -56.72 75 INDIA 897584180 626444530 -271139650 -30.21 76 UNITED KINGDOM 1930723500 1580159540 -350563960 -18.16 77 FRANCE 1710973430 1085437610 -625535820 -36.56 78 NETHERLANDS 2350881380 594655310 -1756226070 -74.71 Author’s own computation 92 University of Ghana http://ugspace.ug.edu.gh Appendix 1. 7: Correlation Matrix EXPORTij GDPi GDPj POPi POPj DISij RBERij AREAj RQij Opennessj POLSTAjt RLjt EXPORTij 1.00 GDPi 0.19 1 .00 GDPj 0.25 0.03 1.00 POPi 0.19 0.99 0.03 1.00 POPj 0.14 0.01 0.43 0.01 1.00 DISij -0.15 0.00 0.30 0.00 0.29 1.00 RBERij 0.25 0.10 0.26 0.09 -0.09 0.16 1 .00 AREAj -0.01 0.00 0.50 0.00 0.43 0.27 0.05 1.00 RQij 0.11 -0.02 0.25 -0.02 -0.09 0.43 0.64 0.02 1.00 Opennessj -0.08 0.03 -0.22 0.02 -0.20 0.23 0.15 -0.26 0.27 1 .00 POLSTAjt 0.02 -0.04 0.10 -0.04 -0.22 0.24 0.49 -0.10 0.75 0.39 1.00 RLjt 0.13 -0.02 0.25 -0.02 -0.07 0.42 0.62 -0.02 0.93 0.27 0.78 1.00 Source: Authors’ calculations Appendix 1. 8: LIST OF COUNTRIES WITH THEIR IMPORT VALUE OF GHANA'S’ NTEs FROM 2004-2016 Year 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Srl Country US$(Million) 1 Angola 0.8 0.4 1.0 0.9 0.7 0.7 0.9 7.9 2.5 6.0 2.9 1.3 0.9 2 Australia 10.8 3.5 7.3 5.8 6.4 3.0 9.2 6.8 8.7 4.9 8.3 6.8 7.4 3 Belgium 20.5 34.9 30.9 29.4 20.3 31.9 66.7 55.0 60.6 54.4 45.6 50.5 69.7 4 Benin 9.6 4.5 8.0 24.9 28.3 27.2 21.5 35.4 38.7 98.4 70.6 32.1 27.3 5 Brazil 0.5 0.6 0.1 2.1 4.8 2.0 4.1 13.3 33.3 19.5 1.8 2.9 3.5 6 Bulgaria 0.2 2.1 0.3 0.1 0.8 1.4 4.2 1.6 0.9 14.7 59.6 69.8 24.7 7 Burkina Faso 16.3 63.3 77.1 70.8 78.8 64.5 83.4 142.7 165.2 197.1 203.9 199.7 248.2 8 Cameroon 1.8 4.4 3.0 1.4 2.7 2.7 4.3 5.8 14.0 6.1 7.7 5.4 5.4 9 Canada 1.9 5.2 5.8 4.5 9.6 11.7 11.1 26.9 6.6 10.3 2.5 9.9 12.7 10 China 8.8 10.8 17.0 14.8 22.3 16.7 22.6 31.7 68.6 42.8 46.3 34.9 34.7 11 Congo 0.3 0.4 2.7 2.7 2.4 2.3 2.5 2.0 1.1 5.4 6.0 0.0 4.9 12 Cote D'ivoire 14.5 18.6 31.8 51.7 52.7 60.4 41.6 59.4 98.3 101.0 102.3 73.5 71.1 13 Denmark 27.8 32.8 9.1 17.9 13.2 15.8 7.1 24.8 23.5 11.6 11.3 27.3 11.2 14 Egypt 4.0 3.5 5.3 7.8 7.9 10.1 11.6 16.6 20.8 18.0 14.7 11.4 14.9 15 Equatorial 0.0 0.0 0.1 0.1 0.3 0.6 1.8 7.7 7.5 5.7 20.1 0.9 0.7 Guinea 16 Estonia 1.4 1.8 3.8 13.2 33.9 18.9 23.5 26.3 44.6 63.8 20.5 27.8 14.8 17 France 80.0 58.1 82.1 131.6 134.1 139.8 182.6 185.6 109.6 108.8 200.9 173.0 124.8 18 Gabon 0.6 1.0 2.4 2.3 4.2 2.0 5.4 4.2 4.3 4.5 7.7 4.6 2.7 93 University of Ghana http://ugspace.ug.edu.gh Appendix 1.8 continued 19 Gambia 1.6 1.2 1.3 2.1 2.0 3.5 4.0 5.1 5.0 3.0 2.3 1.6 2.9 20 Germany 17.8 21.4 45.7 43.3 40.8 36.8 86.7 123.1 79.9 41.3 52.6 88.6 64.2 21 Greece 5.9 3.4 5.1 6.7 4.1 3.0 2.6 4.2 2.9 1.2 1.1 0.5 0.9 22 Guinea 1.5 1.4 1.9 3.0 4.8 5.6 6.8 11.9 9.5 7.4 7.5 9.4 4.7 23 Hong Kong 1.4 1.5 2.2 3.9 3.2 0.9 0.9 1.5 3.5 1.1 0.6 0.6 1.9 24 India 23.6 14.8 47.5 41.3 61.8 37.7 46.0 137.6 104.6 93.7 102.7 102.3 84.2 25 Indonesia 1.7 0.2 0.1 0.3 1.7 1.1 0.3 2.8 2.2 3.5 0.4 2.9 1.6 26 Italy 0.3 0.8 2.3 2.2 1.6 0.4 0.3 1.8 3.0 3.6 5.0 5.9 5.3 27 Japan 7.0 5.4 3.9 8.0 4.6 4.2 6.2 11.3 13.9 20.0 39.0 21.9 18.2 28 S. Korea 1.5 0.8 0.5 2.7 0.0 0.7 1.8 4.7 5.6 6.8 6.9 7.6 14.5 29 Lebanon 0.0 0.0 0.0 0.0 4.3 2.1 3.4 2.6 4.2 4.2 2.4 3.9 3.0 30 Liberia 2.1 2.1 1.8 2.1 2.9 3.6 4.0 9.4 8.6 14.9 13.4 6.5 8.8 31 Malaysia 0.5 1.3 0.7 4.1 9.8 15.7 14.0 33.2 47.2 32.3 61.9 42.5 54.6 32 Mali 4.5 19.5 8.3 5.3 8.0 12.4 20.7 24.8 32.8 42.5 57.7 76.1 122.4 33 Mauritania 0.1 1.7 1.1 1.3 0.7 0.1 0.6 6.1 1.1 8.1 1.3 0.1 0.2 34 Morocco 0.7 0.7 0.6 1.4 0.9 2.9 2.6 1.8 3.5 2.0 1.8 1.3 6.8 35 Netherlands 30.6 30.6 50.2 83.8 120.3 164.3 250.2 403.4 288.7 243.3 291.8 242.7 150.9 36 Niger 7.7 19.2 10.7 15.0 22.2 13.1 14.1 16.3 22.8 37.2 33.6 29.5 46.6 37 Nigeria 28.2 54.1 67.6 113.6 120.6 85.7 96.3 148.5 144.8 149.9 126.4 129.6 142.3 38 Norway 1.1 0.3 0.1 0.9 0.7 0.7 0.4 7.3 0.7 4.0 2.1 3.0 9.4 39 Pakistan 1.5 0.7 0.6 1.2 0.9 0.7 0.5 1.7 1.5 1.8 3.0 4.0 2.2 40 Poland 4.8 3.6 10.1 6.2 10.2 11.0 10.2 37.6 39.6 42.4 27.7 31.9 27.9 41 Israel 0.3 0.8 2.3 2.2 1.6 0.4 0.3 1.8 3.0 3.6 5.0 5.9 0.0 42 Portugal 1.6 6.3 6.3 5.2 6.3 4.0 4.2 3.9 6.0 4.3 9.7 8.8 7.4 43 Russia 0.3 0.3 1.1 3.4 4.2 1.3 2.0 1.3 2.7 0.6 3.4 1.4 2.7 44 Senegal 6.2 6.0 6.8 9.6 11.4 11.5 10.6 13.5 11.8 12.9 10.8 12.4 92.5 45 Sierra Leone 0.0 0.0 1.7 5.5 2.6 3.9 3.9 11.8 6.6 13.7 7.7 4.2 3.6 46 Singapore 1.0 2.0 2.1 4.8 6.5 4.2 3.1 9.6 7.4 22.2 36.2 14.9 18.5 47 South Africa 17.3 6.1 7.0 8.6 12.4 5.1 8.5 25.7 13.2 19.5 13.6 13.8 7.9 48 Spain 38.5 36.6 43.2 51.2 35.0 28.9 60.0 71.2 60.8 89.3 106.9 72.3 108.2 49 Switzerland 11.1 7.2 23.8 17.0 31.2 9.9 11.8 16.2 119.7 121.8 44.3 65.7 36.5 50 Thailand 0.7 2.4 0.3 0.6 1.3 0.3 4.6 1.3 2.7 0.9 1.8 0.7 2.7 51 Togo 21.0 36.2 24.4 60.5 99.4 102.7 118.3 168.1 202.5 161.7 121.0 221.2 146.6 52 Turkey 0.1 0.9 3.8 1.1 4.2 0.9 4.6 7.6 39.5 27.3 37.7 47.5 15.3 53 U A E 0.0 0.0 0.0 5.0 8.9 3.1 6.5 26.5 27.4 17.9 21.1 16.8 16.4 54 U K 111.3 113.4 108.2 133.0 143.4 122.8 130.5 208.7 144.9 163.6 156.5 191.0 203.3 55 U S 83.5 50.0 55.0 62.2 61.2 54.9 107.2 97.7 90.3 93.9 76.9 75.0 113.6 56 Vietnam 1.1 0.6 2.0 2.3 4.9 5.0 6.2 25.0 18.2 61.6 72.3 113.5 154.7 57 Chad 0.0 0.1 0.1 0.2 0.1 1.3 1.1 1.2 0.2 0.5 0.4 1.1 0.3 94 University of Ghana http://ugspace.ug.edu.gh Appendix 1.8 continued 58 Croatia 0.4 0.2 1.5 2.4 1.6 1.0 1.0 0.8 0.4 0.3 0.3 0.2 0.2 59 Finland 1.8 1.3 0.9 2.7 1.4 0.5 2.5 1.1 1.0 1.6 0.3 0.1 0.0 60 Guinea Bissau 0.0 0.5 0.1 0.1 0.6 1.8 1.5 0.5 0.6 0.4 1.0 0.4 0.2 61 Iran 0.0 0.0 0.1 0.0 0.1 0.0 2.1 5.3 2.4 3.9 0.2 0.2 0.8 62 Ireland 1.4 1.5 1.0 1.7 0.5 0.4 0.0 0.6 0.1 2.4 2.3 2.2 1.0 63 Kenya 0.3 0.2 0.3 0.4 0.1 0.2 0.1 0.1 0.7 2.5 1.3 0.5 0.6 64 Peru 0.0 0.0 0.1 1.9 0.0 0.1 0.3 0.5 0.9 0.1 12.5 0.3 0.0 65 Uganda 0.0 0.1 0.0 0.1 0.1 0.0 0.1 0.6 1.5 0.0 0.0 0.0 0.0 66 Sweden 0.7 0.3 0.5 0.4 0.5 0.5 0.2 1.5 1.0 1.5 4.2 5.5 4.3 67 Tunisia 0.1 0.1 0.3 0.0 0.0 0.0 0.6 0.7 0.5 0.2 0.1 0.1 0.0 68 Tanzania 0.1 0.1 0.1 0.0 0.2 0.1 0.2 0.7 0.7 1.6 0.8 1.0 1.8 69 Luxembourg 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 1.1 0.2 0.0 70 Namibia 0.0 1.0 0.0 0.7 0.2 0.0 0.0 0.0 0.2 2.1 0.1 0.0 3.4 71 Botswana 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 72 Mozambique 0.0 0.0 0.0 0.0 0.7 0.0 0.0 7.8 0.1 0.5 0.8 0.0 0.1 73 Bahamas 5.2 10.2 0.1 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0.1 1.3 1.8 74 Cyprus 2.6 1.0 0.0 0.1 0.0 0.0 1.9 0.2 0.1 0.1 0.9 0.0 0.0 75 Panama 1.2 2.1 0.0 0.3 0.3 0.4 0.2 0.0 0.0 0.0 0.0 0.6 0.1 76 Chile 1.0 3.0 0.3 1.9 1.8 0.0 0.0 0.0 0.0 0.0 0.0 0.3 0.1 77 Mexico 0.7 0.6 0.2 0.1 0.0 0.3 1.0 6.9 0.0 0.1 0.0 0.0 1.3 78 Lithuania 0.6 0.1 0.3 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.6 0.0 0.0 Source: GEPC DATABASE 95