University of Ghana http://ugspace.ug.edu.gh THE EFFECT OF CLUSTERING ON PROFITABILITY AMONG SMALL SCALE PALM OIL PROCESSORS IN THE BIRIM NORTH DISTRICT OF GHANA BY EVELYN LARTEBEA GYAMPO (10600194) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE AWARD OF MASTER OF PHILOSOPHY DEGREE IN AGRIBUSINESS DEPARTMENT OF AGRICULTURAL ECONOMICS AND AGRIBUSINESS COLLEGE OF BASIC AND APPLIED SCIENCES SCHOOL OF AGRICULTURE UNIVERSITY OF GHANA, LEGON JULY, 2018 University of Ghana http://ugspace.ug.edu.gh i DECLARATION I, Evelyn Lartebea Gyampo, author of this thesis “THE EFFECT OF CLUSTERING ON PROFITABILITY AMONG PALM OIL PROCESSING ENTERPRISES IN THE BIRIM NORTH DISTRICT” do hereby declare that except for references cited which are appropriately acknowledged, this entire work was done by me in the Department of Agricultural Economics and Agribusiness, University of Ghana. This work has never been presented either in whole or in part for any other degree in this University or elsewhere. Signature………………… Date…………………….. Gyampo, Evelyn Lartebea (Student) This thesis has been submitted for examination with our approval as supervisors: Signature………………… Date……………………. Dr. Henry Anim Somuah (Major Supervisor) Signature………………... Date……………………. Dr. Irene S. Egyir (Co Supervisor) University of Ghana http://ugspace.ug.edu.gh ii DEDICATION I dedicate this work to my parents, Mr. Maxwell Gyampoh and Mrs. Elizabeth Gyampoh, and siblings, Florence, Francis and Eunice. God bless them abundantly. University of Ghana http://ugspace.ug.edu.gh iii ACKNOWLEDGMENTS I thank God, for his wonderful direction, assurance and motivation. His unending kindness and care has brought me this far. I am appreciative to my major supervisor Dr. Henry Anim Somuah for suggesting the study topic and providing guidance. My ardent appreciation goes to. Dr. Irene S. Egyir my co- supervisor for being an excellent and committed supervisor. Her help, resilience, and consolation helped me to finish this work. I feel grateful to all the Lecturers at the Department of Agricultural Economics and Agribusiness for their reactions and attentive help. My significant thanks to all my course mates, particularly Patience Bruku, Eunice Essel, and Nana Ama Darkwaa for their help. Emmanuel Nii Ofei Dodoo likewise merits my significant gratefulness for his priceless counsel all through my scholastic work. Much obliged to you! God bless! I would also like to express my deepest appreciation to the palm oil processors in Birim North District for the opportunity they granted me to interact with them concerning their operations. It was a wonderful experience An expression of thankfulness goes to my parents for their monetary and moral support and all the help during my course of study. To, my siblings, particularly Francis Danquah Gyampo, Mrs. Eunice Agorgey, and Florence Ayeh-Adjei, I say thank you for your help. Evelyn Lartebea Gyampo University of Ghana http://ugspace.ug.edu.gh iv ABSTRACT The study was undertaken to examine the effect of clustering on profitability among small scale palm oil processing enterprises in the Birim North district of Ghana. The study sought to examine the nature of clustering among palm oil processing enterprises, the net income derived from clusters and the extent of cluster influence on profitability among small scale palm oil processors. A multistage sampling was employed to select 200 respondents for the study. Principal Component Analysis (PCA) and K-means cluster analysis was used. PCA was applied on likert scale variables measuring the extent to which small scale palm oil processing enterprises are clustered. Clusters from K-means were used as clusters of palm oil processing in the study area. Gross Margin was employed to estimate the profitability level, and the augmented Cobb Douglas model was employed to determine the extent of cluster influence on profitability among small scale palm oil processing enterprises. Results from the cluster analysis showed that small scale palm oil processors were clustered into on three groups (those that rely on common suppliers, business information, common input material and infrastructure (Cluster 1), those that rely on common suppliers, business information and common market (Cluster 2), and those that rely on transportation and infrastructure (Cluster 3). Palm oil processors in cluster 3 recorded the highest profit of (GHS 11461.28) per month compared to their counterparts in clusters 1 (GHS 6824.59) per month and 2 (GHS 6272.02) per month. The difference among the clusters in terms of their profit level was statistically significant. Results from the augmented Cobb Douglas model also revealed that palm oil processors in cluster 3 had their profit being increased by 0.76% compared to their counterparts in cluster 1. However, the results observed no effect on cluster 2. The study recommended that state policies should encourage private sector participation in setting up training and information service centers within the clusters, palm oil processing firms should be encouraged to provide support for process and product upgrading to improve the skill sets of their employees. The one-district, one factory project by the ruling New Patriotic Party (NPP) government should consider palm oil production in the Birim North district and channel resources to the district in to make it attractive for non-governmental organisation that are willing to invest in the sector. University of Ghana http://ugspace.ug.edu.gh v TABLE OF CONTENTS Content Page DECLARATION i DEDICATION ii ACKNOWLEDGMENTS iii ABSTRACT iv TABLE OF CONTENTS v LIST OF TABLES vii LIST OF FIGURES viii LIST OF ACRONYMS ix CHAPTER ONE 1 INTRODUCTION 1 1.1 Background of the Study 1 1.2 Problem Statement of the Study 3 1.3 Objectives of the Study 6 1.4 Relevance of the Study 6 1.5 Organisation of the Report 7 CHAPTER TWO 8 LITERATURE REVIEW 8 2.1 Introduction 8 2.2 Overview: Palm Oil Evolution in Ghana 8 2.2.1 The situation before and after independence 8 2.2.2 Development of large oil palm plantations in Ghana 13 2.2.3 Cultivation trend of Palm oil 14 2.3 Nature of Small Scale Enterprises 15 2.4 Conceptual Explanations and typology 17 2.4.1 Concept of clustering 17 2.4.2 Typology of the clusters 19 2.4.3 Benefits of clustering 21 2.5 Effect of Clustering on Profitability 22 2.5.1 Indicators of profitability associated with clustering 22 2.5.2 Factors of clustering associated with profitability 23 University of Ghana http://ugspace.ug.edu.gh vi 2.6 Some Empirical Studies Related to the Study 25 CHAPTER THREE 29 METHODOLOGY 29 3.1 Introduction 29 3.2 Theoretical Framework of the Study 29 3.3 Conceptual Framework of the Study 32 3.4 Methods of Data Analyses 33 3.4.1 Describing the nature of clustering of small-scale palm oil processors in the BND 34 3.4.2 Estimating the profitability level among clusters in palm oil processing 36 3.4.3 Determining the effects of cluster influence on profitability of small scale processors 37 3.5 Method of Data Collection 39 3.6 Geographical Area of Study 40 CHAPTER FOUR 44 RESULTS AND DISCUSSIONS 44 4.0 Introduction 44 4.1 Socio-economic Characteristics of Respondents 44 4.2 Background characteristics of the palm oil processing industry 47 4.3 The Nature of Clustering of Small-Scale Palm Oil Processors in BND 48 4.3.1 Principal Component Analysis of the Nature of clustering of Small-Scale Palm oil Processors 52 4.4 The K - means Clustering Results 56 4.5 Profitability of Palm Oil Processing Clusters in the Birim North District 59 4.6 Effect of Cluster Influence on Profitability of Small-Scale Palm Oil Processors 61 CHAPTER FIVE 64 SUMMARY, CONCLUSION AND RECOMMENDATION 64 5.1 Introduction 64 5.2 Summary of the Study 64 5.3 Conclusions of the Study 65 5.4 Recommendations of the Study 66 REFERENCES 67 APPENDIX I 77 University of Ghana http://ugspace.ug.edu.gh vii LIST OF TABLES Table Page Table 3.1 Description of variable for effect of cluster influence on profitability 38 Table 4.1 Socio-economic characteristics of palm oil processors in BND 45 Table 4.2 Background characteristics of palm oil processors in BND 47 Table 4.3 Palm oil processors perception towards having common supplier 49 Table 4.4 Palm oil processors perception towards having common labour 49 Table 4.5 Palm oil processors perception towards having common market 50 Table 4.6 Palm oil processors perception towards transportation 50 Table 4.7 Palm oil processors perception towards storage infrastructure 51 Table 4.8 Palm oil processors perception towards common material 51 Table 4.9 Palm oil processors perception towards sharing business information 52 Table 4.10 Principal component loading estimated scores 53 Table 4.11 KMO and Bartlett’s Test 54 Table 4.12 Rotated components: Varimax with Kaiser Normalization 56 Table 4.13 Participation of processors in combination of nature across clusters 56 Table 4.14 Differences in characteristics among the 3 clusters 58 Table 4.15 Gross margin of palm oil processors among the clusters 60 Table 4.16 Effect of cluster influence on profitability 62 University of Ghana http://ugspace.ug.edu.gh viii LIST OF FIGURES Figure Page Figure 2.1 Palm oil consumption trends from 2011 to 2015 15 Figure 3.1 Conceptual framework: Linking clustering to profitability 33 Figure 3.2 Map of the Study Area 43 Figure 4.1 Scree plot of Eigen values after PCA 55 University of Ghana http://ugspace.ug.edu.gh ix LIST OF ACRONYMS ACs Agro Based Clusters BND Birim North District BOPP Benso Oil Palm Plantation CPO Crude Palm Oil FAO Food and Agriculture Organisation FFB Fresh Fruit Bunch GDP Gross Domestic Product GLSS Ghana Living Standard Survey GM Gross Margin GOPDC Ghana Oil Palm Development Company GSS Ghana Statistical Service ILAPI Institute for Liberty and Policy Innovation ITC International Test Conference MoFA Ministry of Food and Agriculture OLS Ordinary Least Square RGD Registrar General Department SMEs Small and Medium Scale Enterprises SRID Statistics, Research and Information Directorate TC Total Cost TOPP Twifo Praso Oil Palm Plantation TR Total Revenue VC Variable Cost University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE INTRODUCTION 1.1 Background of the Study A key economic activity in the rural areas is Agriculture. The sector also engages about 73.5% agricultural households (Nyanteng and Dzah, 2013; Ghana Living Standard Survey, 2014). Agriculture contributes immensely to the development of the nation, yet it is categorized by low output level. The agricultural sector is estimated to have an average growth rate of 3.3% in the medium term from the 2017 Budget. The Institute for Liberty and Policy Innovation Ghana (ILAPI) (2017) said, “The government cut its 2016 expenditure on the agricultural sector by GHC40 million despite growth in the sector”. Among the several tree crops in the country, oil palm is the next importance to cocoa. In the forest zone, it is one of the prime cash crops. As an essential oilseed, palm oil is used to produce many products which are useful both domestically and can feed our industries in Ghana with raw materials (Ofosu-Budu & Sarpong, 2013). According to the World Bank (2010), palm oil has turned out to be the top vegetable oil manufactured worldwide, responsible for one fourth of worldwide usage and almost 60% of international trade in vegetable oils. Again, palm oil has several competitive advantages over other contending oils, such as high produce, and minimising the high cost of production (Oil World, 2008). In Ghana, the palm oil sector has always been occupied by smallholder farmers and processors who contributed about 93% of total production in 1960. They are still dominant actors in the palm oil industry and are presently responsible for about 80% output of fresh fruit branches and palm oil in the country (MASDAR, 2011). Majority representing 85% of global outputs are produced in Malaysia and Indonesia whiles other smaller but significant producers include 1 University of Ghana http://ugspace.ug.edu.gh Nigeria, Cote D’Ivoire, Columbia, Ghana, Costa Rica, Ecuador, Honduras and Cameroun (World Bank, 2010). Growth in Africa is closely tied to small scale industries. Small scale industries though low paying reduce poverty among the teaming men and women by providing them with jobs (McCormick, 1999). Porter (1998) defines clustering as a geographic concentration of interconnected companies and institutions in a particular field. Cluster as a spatially limited critical mass (that is sufficient to attract specialised services, resources, and suppliers) of companies that have some systematic relationships to one another based on similarities or complementarities (Rosenfeld, 2002). Clustering is mostly thought to give impetus to the positive relationship between small scale enterprises and industrialization as it gives rise to collective efficiency; facilitates growth and creates ease in approaching accessible opportunities and various crises (McCormick, 1999). McCormick (1999) and Mitullah (1999) posits that, the merits of cluster models provided globally can be exemplified in Africa. They recognised that clusters rely on institutional structures, operational systems and the nature of industrialisation. In their study they realised 3 kinds of clusters established in Africa. The groundwork cluster is the foremost, followed by industrialising clusters and lastly the complex cluster. Groundwork cluster establishes the way the processor can get access to their markets. Industrialising clusters on the other hand initiated the systems for differentiation and specialisation and finally the complex cluster which is diversified by its nature or scope and its ability to meet demands of the international community. In Africa, the noticeable one is the groundwork cluster, then industrialising clusters and just minute of complex clusters are found. 2 University of Ghana http://ugspace.ug.edu.gh Pyke and Sengenberger (1992) figured out that the issue of many new businesses is not because they are small but rather by the mannerism in which they are dispersed within the country. The reason being that, these new enterprises do not possess the capacity to give feedback to the pressure that arises from competition and develop a plan for growth. The effect of competition and location of local clusters in the global economy according to Porter (2000), point to the fact that Small and Medium Enterprises (SMEs) in developing nations can progress whereas being competitive in the cluster. Professionals in well-established countries and industrialised entities have exhibited that these Small and Medium Enterprises can grow through clustering. Gebreeyesus and Mohnen (2013) in their study realised that many businesses and knowledge interactions by large appear together simultaneously in clusters. 1.2 Problem Statement of the Study In Ghana, the national Gross Domestic Product (GDP) from agriculture increased to GHS 8,441.00 Million in 2017 from GHS 7790.18 in 2016. The GDP from agriculture in Ghana averaged GHS 6,699.52 million, attaining its peak of GHS 8,441.00 million in 2017; in 2007 it attained a minimum record of GHS 5,322.00 million (Ghana Statistical Service, 2018). In 2016, the Agriculture sector developed from a growth rate of 3% to 8.4% in 2017. In 2017, its share of GDP, however, declined from 18.7% in 2016 to 18.3% in 2017. Crops such as cassava, maize, millet, yam, rice, cocoa and oil palm contribute largely to the agricultural sector with a share of 14.2 % of GDP. (GSS, 2018). Agriculture cannot be disregarded in Ghana’s economic drive, because of its contribution to economic development in the following ways: by providing food and raw materials for non-agricultural sectors of the economy, by creating demand for goods 3 University of Ghana http://ugspace.ug.edu.gh produced in the non-agricultural sectors, by providing investable surplus earned from selling agricultural products in the form of savings and taxes to be invested in non-agricultural sector, earning valuable foreign exchange through the export of agricultural products and by providing employment to a large number of uneducated and unskilled labour. In Ghana, crops can be cultivated in various tropical zones from the rainforest to the dry climate across the ten (10) regions. Tree crops in Ghana include coffee, sheanut, coconut, palm oils, and rubber whereas cash crops also includes maize, wheat, tomato, pineapple, rice and yam. These crops form the foundation of the Ghanaian economy but for this study, the focus is on oil palm. Many have written about palm oil processors in Ghana, (for example, Poku, (2002); Adjei- Nsiah, et al (2002); Owusu, (2007); Gyasi, (2008); Asante, (2012); Osei-Amponsah, et al (2012); Ofosu-Budu and Sarpong, (2013)); Additionally, there have been several works on the concepts of clustering as well as the effects, impact and importance of clusterisation on the development of SMEs in various scientific studies. For instance, Schmitz (2007) study focused on competitive advantages derived from clustering of firms which comes from localised external economies of scale such as skilled labour, better communication and enhanced transportation system. Kuah (2002) sought to bridge competition and networking, just as was reviewed in Porter (1998) work where he used the Clustering theories. Porter explained that the closeness of enterprises in a geographical location, guarantees harmony, increases occurrence and influence of transactions. All these have implications for the small entrepreneurial business and marketing entrepreneurship interface. Adjei-Nsiah et al (2012) in their study on the production practices and profitability of palm oil in Kwaebibirem district observed that, small scale processors are faced with inaccessibility to remunerative market. 4 University of Ghana http://ugspace.ug.edu.gh Ayakwah et al (2018) explained the effects of business structures that have less horizontal competition such as competing retailers or competing wholesalers but with higher vertical cooperation have on clusters. Clustering methods with informal entrenched setting reasonably have higher levels of collaboration yet minimal competition in relationships. The study went on to elaborate that inter-business relationships have a good effect on the operations of clustered entities in terms of sharing of innovative ideas, drafting a contract, availability of business funds for operations, and the way they collaborate to various interest groups within their activities. Soyebo et al (2005) in their study, encountered challenges and limitations influencing palm oil development in Ife of Osun State, Nigeria. The study established the proportion of tree crop growers that actively participated in the cultivation of palm oil. They encouraged smallholder farmers to get involved in cooperatives. Also, with regards to the Birim North District, emphasis on the problems and constraints to the growth of palm oil processors is key. However, minimum levels of technology for processing, small-sized markets for products, inadequate access to capital, insufficient physical infrastructure and support and the available organisational framework affects the profitability of palm oil processors and the fact that they have no knowledge on clustering and the benefits it brings to them. Hence the following research questions become pertinent: 1. What is the nature of clustering of small-scale palm oil processors in Birim North District? 2. What is the profitability level among the palm oil processors? and 3. What is the extent of cluster influence on the profitability of small-scale palm oil processors? 5 University of Ghana http://ugspace.ug.edu.gh 1.3 Objectives of the Study The major objective of the study is to access the effect of clustering on profitability among small-scale palm oil processing enterprises in the Birim North District of Ghana. The specific objectives are to: 1. Examine the nature of clustering of small-scale palm oil processors, 2. Estimate the profitability level of clusters in palm oil processing and 3. Determine the extent of cluster influence on the profitability of palm oil processors. 1.4 Relevance of the Study Many studies have been conducted in Asia and West Africa into palm oil processing and its benefits to the continents because of its increased demand worldwide (Corley and Tinker 2008). The research explores the inner workings of clustering in the small-scale palm oil processing industry. The paucity of research on the effect of clustering on the profitability of palm oil processors in Ghana is a key driver of this study. Findings from the nature of clustering will add-on literature to earlier research works. It will also serve as a guide for students and other researchers in this discipline. Furthermore, in a country such as Ghana where resources are limited in nature and opportunities for information technology is inadequate, findings from the estimation of profit among clusters will serve as a possible way of cluster formation for processors to generate more profit. Politically, it will serve as an advisory tool for governmental and non-governmental institutions in policy formulation. Findings from the effect of clustering influence on profitability will help management of palm oil processing and other certification bodies to know where they are lacking in the industry and form a baseline for subsequent studies 6 University of Ghana http://ugspace.ug.edu.gh about the clustering of small-scale palm oil processing to enhance knowledge and improve policy interventions for the survival of small-scale palm oil processing. 1.5 Organisation of the Report For a meaningful and coherent presentation, the thesis has been structured into five chapters. The first chapter comprises the introduction of the study. It also indicates the problem statement followed by research questions, objectives of the study, and finally the relevance of the study. Chapter two constitutes a review of studies by other researcher that are related to the study. Chapter three describes the research methodology involving theoretical framework, conceptual framework, method of data analysis and data collection as well as geographical area of study. Chapter four presents and discusses the results obtained from the study. Finally, Chapter five presents the summary, conclusions, and recommendations of the study. 7 University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO LITERATURE REVIEW 2.1 Introduction This chapter presents summaries of the literature reviewed. First pre and post-independence initiative of oil palm production and oil palm cultivation in Ghana is described. Then, the nature of small scale enterprises and typology of clusters and benefits of clustering follow. Literature on the relationship between clustering and profitability is then presented. Finally, some empirical studies related to the study are also presented. 2.2 Overview: Palm Oil Evolution in Ghana 2.2.1 The situation before and after independence The situation before independence: Aghalino (2010) observed that the indigenes of West Africa and then Gold Coast had managed their economy by putting a system in place mainly for oil palm production far before the arrival of the Europeans. The earliest archaeological evidence exposed oil palm consumption located in an Egyptian tomb in Abydos, but it was identified that the palm oil consumed in Egypt was exported from West Africa to North Africa (MOPB, 2009). In 1820, Ghana’s main international profitable trade in palm oil happened with wild palm harvesting. “Oil palm evolved into cash crop and plantations were established by 1850. This led to palm oil becoming the principal export from the then Gold Coast (MoFA, 2011). Consequently, palm oil trade became important in the second half of the 19th century after the abolition of the slave trade railway (Dickson, 2009; Lynn, 2002; Bergert, 2000, MAFAP, 2013). These developments encouraged indigenes in the southern sector of the Gold Coast to cultivate cash crops for the international 8 University of Ghana http://ugspace.ug.edu.gh trade. The colonial administration also played a role in this instance, by promulgating agriculture programs. This is because agriculture was a way to increase export of cash crops to Europe (Dickson, 2009). As a result, trade in all forms of agricultural produce began to blossom especially, palm oil. During this period, palm oil was harvested from the wild groves. Gyasi (2012) asserts that the oil palm grew in the wild in some areas located near the coast. Specifically, the oil palm belt in the wild extended for about 500km from the southeast to the southwest coast with an approximate width ranging from 8 to 60 km. As palm oil gained importance in the international trade, people began to cultivate the oil palm within the same belt which was colonized by wild palmeries (Gyasi, 2012). According to Gyasi (1992), palm oil trade began to flourish in the pre-colonial era. European developed oil palm plantation along the coast, besides those obtained from palm groves in the wild to produce palm oil for the European markets. The colonial government at that time believed that the local small-scale farming operation was stronger economically than foreign plantations and, as a result, plantations did not make much impact on agricultural production during the colonial period in Ghana. The situation after independence: In 1960 the government of Ghana began to put in much effort to encourage the palm oil industry (Danyo, 2013). Danyo (2013) said, “This was directed by various Government policies and programmes that were formulated and published in eight (8) national development plans and strategy documents. The Ghana Palm Oil Research Institute at Kusi, was established and vested with the mandate of producing high yielding oil palm varieties in Ghana” (MOFA, 2011). State 9 University of Ghana http://ugspace.ug.edu.gh farms and nucleus estates cultivated with oil palm, and processing mills were also established at various part of southern Ghana. These initiatives contributed to the growth and spread of oil palm in new areas, north of the old belt (Gyasi, 2008; Gyasi, 2012). In addition, urbanisation and change in climatic conditions along the coast had also caused a decline in the cultivation of oil palm in the old belt. Subsequently, there was a shift on the production of oil palm to its ecological niche (Gyasi 2011). According to Gyasi (2012), the largest oil palm growing areas in the new oil palm belt extended westward from the Lake Volta to Benso in the Western region. Also, the area around Kade represents a core region of oil palm. Oil palm extends to about 100 km from Asamankese through Kusi, Kade, and Kwae to Akoase near Nkawkaw in the Eastern region (Gyasi, 2012). The vast expansion in oil palm plantation under a five-year development plan (1959- 64) was meant to introduce economic and industrial development in the country (Ntsiful, 2010). Ghana Farmers Council distributed oil palm seedlings to its members under this program. In addition, 2, 565 hectares of oil palm were cultivated in various parts of the country with a palm oil processing mill established at Asesewa (MASDAR, 2011). Despite the promising nature of the policy, state farm projects were not fruitful because of mismanagement, lack of financial support and political interference through 1966 military coup (Ntsiful, 2010). Two different administrations that is, the National Liberation Council and the Progress Party (1966-1972) tried to revitalise the palm oil industry by privatizing some estates but they were unsuccessful because of the political situation prevailing in the country (Foli 2010 cited in Fold & Whitfield, 2012). The palm oil industry was fully revived under Acheampong’s 10 University of Ghana http://ugspace.ug.edu.gh administration after a five-year development plan was instituted in 1975/76. This was because domestic production of oil palm in 1972 was meeting only 43% of demand meant for industrial processing (Fold & Whitfield, 2012). The new policy called for private sector support through financial and technical assistance to smallholder and out-grower farmers. This was an integral part of plantation development and the state was a lead player. Subsequently, the state received external support from Commonwealth Development Corporation, European Community, World Bank, International Finance Corporation, African Development Bank, FMO (Dutch Financing Company) and Agence Française de Development. Through their support, Benso Palm Oil Plantation (BOPP), Twifo Praso Palm Oil Plantation (TOPP), and Ghana Palm Oil Development Corporation (GOPDC) were established. However, the existing National Palm Oil Limited (NOPL) was revamped with support from external donors and the State. Value addition was carried out in the palm oil industry in a different dimension during structural adjustment era (1980‟s to early 1990s). As part of the five-year development plan, there was privatization of non-viable state properties including state farms and processing mills. The expropriation of land during the creation of plantation projects in the 1960s and 1970s caused land litigations, protests by peasant farmers and delays in getting the large estates operational during the privatization exercise (Gyasi, 1992b; Gyasi, 1996). Privatization of state owned palm oil plantations and estates continued in the late 1990s. In 1994, the government of Ghana sold 80 percent of its shares and management control in GOPDC to SIAT Ghana Consortium. 11 University of Ghana http://ugspace.ug.edu.gh In 2004, the state offloaded 58% of its shares in BOPP to Unilever. In 2008, 20% of the state’s shares in BOPP were also sold. In 1997, government sold 40 percent of its shares in TOPP to Unilever. Therefore, BOPP and TOPP are now managed jointly by Unilever. The National Oil Palm Limited was sold in 2000 to NORPALM to a company from Norway. Small state-owned plantations and mills were bought by Ghanaian entrepreneurs or by large estates with the aim of expanding palm oil production (MASDAR, 2011; Fold & Whitfield, 2012). As a result of privatisation of state owned plantations, there were no palm oil farms owned by the state. In 2003, a new palm oil policy which was a special presidential initiative (PSI) on palm oil was introduced to recuperate and expand cultivation and processing. PSI on the palm oil aimed at building the rural industry, creating rural employment and empowering smallholder farmers (Asante, 2012). Under the initiative, landowners contributed land within a certain radius to be used by groups of farmers. The total area of palm oil cultivated under the initiative, expanded by 10,000 hectares of small scale farms with high-yielding varieties. The scheme later failed because of past experience of state in expropriating communal land, political apathy, insufficient funds, and poor management (Fold &Whitfield 2012; Asante, 2012). Afterward, a new palm oil master plan was launched in July 2012 under a different political administration. The plan is designed to guide the state to promote a competitive palm oil sector in the country. It proposed a 10,000-hectare palm oil nucleus estate and processing mill to be sited in the Prestea-Huni District in the Western Region. It also included a village level small mill which is to be associated with a proposed nucleus estate (MASDAR, 2011; Bonney, 2012). With the master plan, it is expected that Ghana will experience a growth in the total cultivated area under palm oil (MASDAR, 2011). The contribution of international donors, state 12 University of Ghana http://ugspace.ug.edu.gh motivations, and policies are forms of support services and enabling environment created to enhance value addition in the oil palm industry. 2.2.2 Development of large oil palm plantations in Ghana The development of large palm oil plantation started in the 1970s (Huddleston and Tont, 2007). According to Huddleston and Tont( 2007) the government during the period shifted its focus on agriculture to privatisation of large-scale plantation. The socio-economic importance of conventional agriculture was also recognized, and the government ensured that contract farming is made a core part of any agribusiness operation (Daddieh, 1994). The government- sponsored scheme, called for the private sector to support financially and technically to either smallholders who are contracted tenant farmers on the nucleus plantation estate or independent farmers who sharecrop, rent or own their land outside the confines of the estate (Huddleston and Tont, 2007). According to Carrere (2010), there are three main oil palm plantations, namely; Ghana Palm Oil Development Co. (GOPDC) located in Kwae (Eastern Region); Twifo Palm oil Plantations Ltd. (TOPP) situated in Twifo Praso/Ntafrewaso (Central Region); and The Benso Palm oil Plantations Ltd. (BOPP) located around Benso/Adum Banso (Western Region). They were supposed to produce oil from the palm fruit they planted (Gyasi, 1992). These companies encouraged palm fruit production among the peasant farmers in the plantation to help sustain their huge palm-oil-processing mills located inside the plantations through the nucleus estate system in an attempt to develop the acquired areas into palm plantations Carrere (2010). 13 University of Ghana http://ugspace.ug.edu.gh In the early 1900s, cocoa began to flourish and quickly surpassed oil palm as the primary agricultural cash crop (Gyasi, 1992), but after the three plantations have been developed, they contributed meaningfully towards the growth of Ghana‘s oil-palm hectares from 18,000 to 103,000 between 1970 and 1990 (Carrere, 2010). This growth (24% per annum) has resulted in the re-emergence of the palm oil as a major commercial crop and has served as a basis for the fast-developing palm oil and other agro-industrial industries and granted the country more than self-sufficient in palm oil production (Fold, 2008). 2.2.3 Cultivation trend of Palm oil Poku (2002) in his study said palm oil plantations in Ghana occupied 304,000ha. In 2004, 285,000 hectares of palm oil were cultivated (Carrere, 2010). Cultivation by smallholders nearly covered 88% of the total land space for production but produced only 72% of the palm oil fresh fruit bunch (FFB). The remaining 28% was produced by private businesses cultivating fewer than 12% of the total area. The existing plantations operate based on a nucleus estate with associated smallholder schemes and independent out-growers. According to MoFA (2011), total land under palm oil was given as 360,000 hectares as of 2010. Most of the palm oil produced (see figure 2.1) in Ghana for human consumption is processed at the small-scale level. Ghana has an estimated 336,000ha of land under palm oil cultivation (MASDAR, 2012) and produces about 243,852 tons of palm oil. Fold (2008) said the amount of palm oil produced in Ghana is not known as a result of a substantial amount of home production that is purposely consumed at home, whereas for industrial use they are operated by foreign companies (BOPP, TOPP, GOPDC, and NORPALM). 14 University of Ghana http://ugspace.ug.edu.gh 1000 800 760 730 700 665 600 400 200 0 2011/2012 2012/2013 2013/2014 2014/2015 Figure 2.1 Palm oil consumption trends from 2011 to 2015 Source: (GSS, 2015) Figure 2.1 shows the palm oil consumption pattern in Ghana from 2011/2012 to 2014/2015. In 2014/2015, the consumption of palm oil in Ghana was on its peak with a consumption of 760,000 metric tonnes representing an increase of 3.95% from 2013/2014. In 2013/2014 there was a consumption of 730,000 metric tonnes which indicated an increase of 4.11% from the previous year. There was a 5% increase in 2012/2013 with a consumption of 700,000 metric tonnes. By 2011/2012 the consumption for the period was 665,000 metric tonnes. (GSS 2015). 2.3 Nature of Small Scale Enterprises Small and medium scale enterprises (SMEs) are lifeblood of most economies. Abor and Quartey (2010) discussed the characteristics and contributions of SMEs to economic development. They indicated that SMEs in Ghana provides about 85% of manufacturing employment in Ghana. 15 consumption in thousand metric tonnes University of Ghana http://ugspace.ug.edu.gh Snodgrass and Biggs (1996) argued that small firms employ many labour forces in many developing countries, but they are more labour demanding than large firm however, many analysts argue that, within industries, SMEs are more labour intensive than large firms. On the average SMEs represent over 90% of the enterprises and account for 50 to 60% of employment in most African countries. According to Abor & Quartey (2010), SMEs in Ghana have been noted to provide about 85% of manufacturing employment in Ghana. If all stakeholders are to show serious commitment to the development of the SMEs sub-sector, it follows that the economy must necessarily witness meaningful transformation and prosperity. SMEs are described as efficient and prolific job creator, the seed of big businesses and the fuel of national economic engine (Abor & Quartey, 2010). In Ghana, available data from the Registrar General’s Department indicate that 90% of the companies registered are micro, small and medium enterprises. These target groups have been identified as catalyst for economic growth of the country as they are major sources of income and employment (Mensah, 2004). Aside from providing opportunities for employment generation, SMEs help to provide effective means of curtailing rural-urban migration and resource utilization. By largely producing intermediate products for use in large–scale companies, SMEs contribute to the strengthening of industrial inter-linkages and integration. Gunu (2004) and Aremu (2010) both found that Small-Scale Enterprises help in providing income generation, encourage savings, and create employment. They are veritable engines for the development of entrepreneurial capabilities and indigenous technology which generate employment in the country. Small and medium scale industries constitute the basis for industry and natural economy in many countries. It has been estimated that SME’s employ 22% of the adult population in developing countries. They can be regarded as one of the important element 16 University of Ghana http://ugspace.ug.edu.gh of a country development and this play a crucial role in the economy of this nation Ayanda & Laraba (2011). 2.4 Conceptual Explanations and typology 2.4.1 Concept of clustering A lot of interest has been given to the concept of ‘clusters’ by policy drivers in the past. Global clusters have been considered possible drivers of competitiveness and productivity. A nation’s economy entails clusters of industries related through vertical and horizontal relationships (Porter, 1990). Cluster initiatives are usually regarded to be efficient in policy instruments such that they give way for a focus of resources and funding in targeted areas with a high development potential that can spread beyond the target locations. With regards to Porter (1998), Clusters are concentrations in a geographical location of interconnected companies and institutions in a particular field. Clusters can be an array of linked industries and other entities important to productivity. Clusters channel customers to manufacturers of harmonising products and to industries related by technologies or skills, or similar inputs. Most clusters include governmental and other institutions, such as standard-setting agencies, vocational training providers, universities, think tanks, and trade union groups that provide special training, information, education, research and technical support. Clusters also facilitate both competition, profitability and productivity (Sunil & Anupam, 2014). Opponents compete strongly to retain clients. Yet there is collaboration, much of it upright, involving companies in the same local institution. 17 University of Ghana http://ugspace.ug.edu.gh In addition, clustering can be classified under either formal or informal clusters can be found in the public or private sector; and can also be in the form of horizontal or vertical clusters. Again, it can be either physical or virtual. Vertical networks consist of horizontal cluster participants along with supply chain associates such as consumers, suppliers, and related services (Boekholt, 1997). Diagonal clustering refers to the concentration of complementary or symbiotic activities, whereby each firm adds value to the other. Clustering is also partly determined by knowledge diffusion, which relies on two critical factors: (1) geographic proximity, and (2) social structure (Enright 2001). Rosenfeld (1997) distinguishes clustering activities by the intensity of social infrastructure and firm interaction, firmly placing social capital and trust as the basis of collaboration, information and knowledge flows in regional clusters. Swann et al (1998) similarly positions relational capital at the core of cluster strength and as the foundation of its knowledge base. Porter (1998) suggests that a gap exists in literature on cluster around social structures which may be very beneficial for SMEs than what prevails in a mature cluster. Literally, the term cluster has many connotations which can be referred to as a group of firms engaged in similar or related activities within a national economy (Porter, 1990). Schmitz (1992) on the other hand distinct cluster as a group of enterprises belonging to the same sector and operating in close proximity to each other. Schmitz's approach was drawn not only from Marshall's (1890) analysis on economic activities in textile and metalworking districts of England, Germany and France but also from his experience on growth constraints of small industries in developing countries. Porter (1998) further defined clusters as the geographical concentration of interconnected firms and institutions in a particular sector. It is worthy to note that it is vital to strengthen the link that exists between the firms. 18 University of Ghana http://ugspace.ug.edu.gh Complementarity is another terminology used to explain cluster. Steiner and Hartmann (1998) indicated that "Clusters are sets of complementary firms (in production and service sectors) public, private and semi-public research and development institutions, which are interconnected by the labour market and/or input-output and/or technological links". Steiner and Hartmann (1998) further argued that such clusters are highly competitive because the links are governed by the advantages of market mechanism and the direct structures of the single organisation. According to Elsner (2000), clusters are a group of firms that are functionally interconnected vertically as well as horizontal. The functionality connection emphasises the nature of the relationships between firms and the institutions supporting the cluster and such relations according to Elsner are defined through the market. 2.4.2 Typology of the clusters Organised cluster: Entrepreneurship development has four main development trajectories. Consecutively, these are trading, private savings, private investment, and manufacturing. Further investigation revealed that specialisation in specific branches of trade has played an important role in the SMEs development. Moreover, related to the industrial development it was tried to answer some questions, based on empirical enquiries advanced by Komlosy et al. (1997) and Dike (1997), on the relationship between capital accumulation and the informal sector. Agro-based clusters: In developing countries today, the greatest opportunity for reasonable growth lies in the agricultural sector. Yet ironically, it is this sector where poverty is most widespread and found in its worst forms. Small-scale farmers, and the rural communities in which they live, are 19 University of Ghana http://ugspace.ug.edu.gh imprisoned within a “cycle of equilibrium” of low margins, resulting in low risk-taking ability and low investment, which leads to low productivity, low market orientation and low value addition which, in turn, nets low margins (International Test Conference, 2006). From a conceptual point of view, the creation of “value networks” is the most effective means by which to break this cycle, while at the same time raising prospects for long-term competitiveness in the sector of agricultural. In this context, a value network is firstly an aggregation of vertical relationships amongst sellers of raw materials and production inputs, agricultural manufacturers, processors and exporters, branded buyers and retailers (ITC, 2006). There are, of course, exceptions. Greater integration of the value chain is being achieved through vertical relationships that improve product flow (contract farming and out-grower schemes), coordinate financing and payments (ITC, 2006) and reinforce communication. And certainly, cooperatives, joint export marketing and similar efforts to nurture horizontal relationships, are having an impact on rural competitiveness and well-being. In the examples of “higher return relationships” lacks the preferred value network where there should be a continuously closer coordination and collaboration horizontally, vertically and along the support dimension of the value chain and where such relations not only create effectiveness, efficiencies, synergies and opportunities (both developmental and commercial), but also inspire innovation, value addition, and product diversification at the level of small-scale producers and agribusiness, and make sure that optimum gains are reserved in the local economy (ITC, 2006) . One-village-one-product: This is a program that was started in Japan with the aim of promoting regional development. People living in smaller towns, specifically villages, are encouraged to a local product that 20 University of Ghana http://ugspace.ug.edu.gh benefits the society at large with marketing support from government. The products that are produced from this program sold to consumers in the local market and internationally. Initially, Japan seemed to be the only country that spearheaded this movement. However, after years of a very high success rate, the campaign is now part of Japan’s foreign assistance programmes. It is also being used by other countries. Thailand, for example, now has a “One-Tambon-One- Product” scheme (FAO, 2004). Spice export villages in Sri Lanka are another example of agglomeration and facilities provided in specific locations. 2.4.3 Benefits of clustering According to Porter (2000), the occurrence of clusters exposes essential insights about the microeconomics of rivalry and the duty of location in competitive advantage. Porter (2000) had said that, even as old reasons for clustering have diminished in importance with globalization. Scholars like Keeble & Wilkinson (2000), Storper (1997) and others argued why small-scale cluster, providing to them transaction cost savings alone is inadequate to explain the growth and perseverance of clusters. A description is that larger businesses internalise much of the lateral, horizontal and vertical scope of a cluster. Small-scale entities are disadvantaged in their quest to specialised resources and intelligent capital. Taylor and McRae-Williams (2005) suggest that clustering simulates large firm behaviour such that when small firms are not in a position to internalise externalities through economies of scale, they cluster to access resources, to reduce costs, to compete with larger firms, and to innovate. In other words, by networking, firms can share knowledge, small-scale entities are able to compete for and admit specialised resources and information systems. They can as well internalise competencies and assets that typically are internalised by larger firms with economies of scale (Tayler & McRae-Williams, 2005). 21 University of Ghana http://ugspace.ug.edu.gh Clustering then provides Small-scale entities assistance that would be inaccessible or be accessible at a higher cost to non-clustering clients. While value-added and activities such as access to a global client base and advanced business services/production are clearly major contributing factors for small business clustering, the need for access to localised explicit and tacit knowledge networks has proven to be a central driver for clustering (Keeble, 2000). Rubzen et al (2013) recorded the benefits of clusters as an improvement in productivity as well as farmer income. In addition, Rubzen et al (2013) proposed spillover outcome in the kind of improved farm employment. Clustered businesses can improve the route of policy through group action. Felzensztein and Gimmon (2009) informed that improved status or genuineness of the firms and products, purchasing transitional goods from other firms, provision of access to better equiped suppliers, and finding prospective clients in new markets as the main externalities that the respondent’s firms found more useful as benefits of geographical co- location. Farmers and small-scale agricultural business can gain from partaking in Agro based Clusters as it paves way for them to acquire scale economies and shared costs in training, information sharing, certification and technology application. 2.5 Effect of Clustering on Profitability 2.5.1 Indicators of profitability associated with clustering Evidence from rural industrial and agro business clusters in the United States shows that clusters contribute positively and immensely to sectorial economic growth and development as well as appreciable income for the indigenous working class (Henry and Drabenstott 1996; Gibbs and Bernat 1997). Because of their accessibility, cluster constituents and participants enjoy the economic benefits of several facets with specified locations and synergies. A well-developed 22 University of Ghana http://ugspace.ug.edu.gh aggregation of related agribusiness prods three important activities: thus, increased productivity, more rapid innovation and new business formation. However, the various pointers can be categorically defined as follows; increased productivity can effectively be achieved through specialized inputs, access to information, synergies, and access to various trade markets (Bigsten and Tengstam, 2011). More rapid innovation comes to play and developed through cooperative research initiation and competitive venturing. Nothing sparks productivity and innovation better than competition in proximity whether technological tools and skills innovation, as seen in the IT clusters as adopted in the most countries in Asia, or creative innovation, as in the fashion design clusters in the United States and Europe (Aremu, 2010) New business formation can be propelled and developed by filling in niches and expanding the frontiers of the cluster boundaries. In development and clustering, as a result of competition, demand for services, and the attraction of investors; this dynamism in turn intensify novelty (Abor and Quartey, 2010). 2.5.2 Factors of clustering associated with profitability Clustering ensue certain privileges due to the interconnectivity and the synergies that comes with it. These accrue range of benefits so far notable are better and more efficient access to infrastructure, specialized human resources, and inputs, including capital (Ali and Peerling, 2011). Firms readily obtain access to paramount and novel inputs such as suppliers, information, technology, financial aids, skills and institutions of higher education. This readily accessible parameter as a result of the clustering leads to: 23 University of Ghana http://ugspace.ug.edu.gh Reduction of costs: Transaction expenditure are drastically lowered because of the closeness and accessibly in the cluster. Proximity offers crucial benefits for the agro-industries and businesses in developing countries, in particular for SMEs. Often businesses can contract for products and services from inside the cluster and enjoy the benefits or do away with the cost of having to manufacture or develop the product or service. Various expenditure regarded as costs are also being reduced through economies of scale and scope, as in the case of partnership trading and procurements. Costs related to hiring skilled and experience labour is reduced drastically, provided such skills are made readily in the cluster (Sunil and Anupam, 2014). Access to information and services: Participating in a cluster presents participants with preferred access to the global market, technical, and competitive information that is gathered in the cluster. For instance, through a close relationship with worldly-wise and reliable buyers within a cluster, suppliers are more attuned to their specific needs. Business organizations and also business fairs thrive on information and service hubs; informal periodic contact with similar companies also plays an equally significant role the accumulation of these information and service hub (Gibbs and Bernat, 1996). Attraction of foreign investment: If clusters are represented as central drive for their firms and industries, they will allure all the magnates and business mogul from far and near. Actually, exotic firms and industries can enhance the executives of the cluster and contribute to its upgrading, as experienced in the Ugandan fish cluster where this function was played by European firms as well as the South African wine cluster and the Austrian companies (Fafchamps and El Hamine, 2004). 24 University of Ghana http://ugspace.ug.edu.gh Better recognition and marketing: Infant and uprising firms and industries, establishing in a cluster provides them with the opportunity near competitors and firms of similar products or trade identity may help them to advance, gain recognition, and acquire strong footing more rapidly within the trade. The South African wine cluster is indicative (Schwab et al, 2002). This was attained only when firms were established to market wine for a pool of producers did they gain recognition to export and trade at the global level. An individual manufacturer or producer alone could not have made this possible. A cluster often provides the platform to seed or enlighten others as activities are disseminated in the value chain to lessen the risk, access discounted or low-cost inputs, or alternative serves particular regional markets (Abiola, 2006). Export oriented clusters usually generate above-average returns, productivity, and rapid innovation. The South African wine cluster has demonstrated these merits and shows how the development of a cluster can spur innovation and economic growth and development in firms and industries (Oliveira-Wilk & Evaldo, 2003) 2.6 Some Empirical Studies Related to the Study Olagunju (2008) studied the economics of palm oil processing in South-Western Nigeria. He analysed the profitability using multiple regression and hydraulic hand press technique. He concluded that, palm oil processing is economically viable and that, cost of extraction of palm oil was negative yet very importantly it was linked to net return while tools depreciated and other inputs demonstrated negative relationship with net return. No significant relationship was identified to have occurred between net return and other elements such as the experience needed for processing and the labour cost. 25 University of Ghana http://ugspace.ug.edu.gh Navickas and Malakauskaite (2009) analysed the impact of clusterisation on development of SME sector. In their study, Navickas and Malakauskaite (2009) said “the cooperation of companies at the national level and on global scale is becoming more and more important as a tool of economic development”. Navickas and Malakauskaite (2009) realised that these firms tend to work jointly so far as they can share their competencies, reduce various costs, consolidate limited and hereby maximise the output level, innovativeness and profitability. a) Companies tend to cooperate in order to achieve the effect of synergy in various fields (e.g. agro-processing firms) of operation and improve their performance in the competitive environment. b) SMEs that partake in clusters can benefit from specialised infrastructure, increased possibilities to penetrate new markets, qualified workforce, and ability to meet the needs of clients, and cost reduction in manufacturing operations. Adjei-Nsiah et al (2012) in their study on the production practices and profitability of palm oil production among small scale processors in the Kwaebibirem district observed three groups of processing equipment known as Digester Screw press combined, digester with a separated hand operated hydraulic press and digester with separate hand operated screw press. Adjei-Nsiah et al (2012) also observed small scale processors in the industry were faced by inaccessibility to remunerative market especially during the highest fruit production season (Feb. to May). They also found out that, the processors were faced with lack of credit and knowledge. Adjei-Nsiah et al (2012) also realised that in the peak season (April to May), palm oil processors were likely to make a loss of 38% of every cedi in revenue and that palm oil is more profitable during the lean period where oil is scarce. 26 University of Ghana http://ugspace.ug.edu.gh Onoja et al (2012) examined the profitability of cocoa farmers in Ondo state (Nigeria). Onoja et al (2012) employed Net farm Income to estimate the profitability of cocoa production. Their results showed that, cocoa production was profitable with mean profit of 1,634,182.72 naira. Nyunza and Mwakaje (2012) analysed the factors influencing crop profitability of round potato marketing in Tanzania. Ordinary least square (OLS) was employed using primary data from 120 farmers through random selection to determine the factors influencing crop profitability. The results showed that household size and selling price were significant and both had a positive relationship with crop profitability. The results also indicated that education and land size were not significant however they had a positive effect on crop profitability. Umar (2014), employed three functional models (linear, Semi- log and Cobb-Douglas functional model) to estimate the factors influencing profitability among gum arabic marketers in Jalingo Local Government of Taraba state in Nigeria. Fifty respondents were randomly selected and used for the study. The results from the study revealed that semi-log functional form was found to have the best fit based on statistical significance of their coefficient of multiple determinations (R-squared), the magnitude of the standard errors, and the coefficients signs and levels significance. Results from the semi-log functional model showed that family size, purchasing cost, quantity of gum Arabic sold, labour cost and marketing experience of gum Arabic had positive significant effect on profitability. Kanyua et al (2015) also in his study to determine the factors influencing profitability of diversified cash crop farming among smallholder tea farmers in Gatanga District of Kenya. 27 University of Ghana http://ugspace.ug.edu.gh The regression results revealed that gender, farming experience, farm tools, farm size, credit, hired labour and the fertilizer and manure applied were the significant determinants. Foghani et al (2017) employed the merits of cluster-based systems to enable SMEs gain global status. The main aim of their study was to identify the relationship between the capabilities of networks and clusters in developing SMEs preparedness in facing economic players in the global environment. Foghani et al (2017) found out that cluster-based SMEs have the potential to include SMEs in the growth of productivity and profitability and network global distribution channels adequately. In conclusion, Foghani et al (2017) discussed best practices for different forms of economies whether developing, developed and one in transition. They noted that sustainable and growing SME sector can only be achieved if proper support structures are put in place for SMEs as a key business environment. Such support services include Business advice, providing or sharing vital information, training and financial services. 28 University of Ghana http://ugspace.ug.edu.gh CHAPTER THREE METHODOLOGY 3.1 Introduction This chapter comprises of three sub-sections. First and foremost, the theoretical and conceptual frameworks, and then the data analysis of each of the objectives are described. Finally, the method of data collection including sampling procedure and study area are described. 3.2 Theoretical Framework of the Study This study was built on the theory of agglomeration. According Siba and Söderbom (2015), mechanisms around the theory of agglomeration are knowledge spillovers and externalities. Marshall (1890) revealed that mass production, availability of specialized input, proximity to labour supply and availability of modern infrastructure are the externalities relevant for the formation of cluster. This has an effect on production externality which might result in knowledge spillovers. Given that firms are close to each other in geographical location, employees in different firms have informal contact with each other and such result in high turnover in local employees. Firms in this area gain knowledge owned by its workers who accumulate such knowledge from their previous jobs. According to LaFountain (2005), production externalities may also be classified as localization externalities and urbanization externalities. Localization externality occurs when firms benefit from proximity to similar firms or other firms in the same industry and the latter exist when firms benefit from locating in a diverse location or from proximity to firms engaged in variety of different industries. This study was based on the concept of localization externality. Firms located in a location with higher agglomeration will benefit from information spilling over from other firms, reduction in 29 University of Ghana http://ugspace.ug.edu.gh transport costs of inter firm trade, and enhance the diversity of firms and local products available, and hence minimize their cost of production and increase productivity as well. As such palm oil processors are able to maximize their profit through their maximising output from a given resource or minimising the resources required for a given output. From the theory of firm behaviour, palm oil processors derive their profit function from the use of inputs such as palm nuts, water for the production of output (palm oil). Consider a number of firms i= 1, 2, 3,….., N which supply output to the market in their quest to maximize their profit. 𝑃𝑟𝑜𝑓𝑖𝑡 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 = 𝑓(𝑄, 𝑃, 𝐶) (3.1) Where Q denotes Quantity, P denotes Price and C denotes cost of production. 𝑃𝑟𝑜𝑓𝑖𝑡 (𝜋) = 𝑇𝑅 – 𝑇𝐶 (3.2) Where TR denotes Total Revenue and TC denotes Total Cost, and 𝑇𝑅 = 𝑃𝑄 (3.3) The profit maximisation behaviour of the palm oil processor can be expressed as: 𝜋 = 𝑃𝑄 − 𝑇𝐶 (3.4) For the processor to maximise profit, the first order condition must be satisfied. 𝑑𝜋 𝑑𝑇𝑅 𝑑𝑇𝐶 = − = 0 (3.5) 𝑑𝑄 𝑑𝑄 𝑑𝑄 𝑑𝑇𝑅 𝑑𝑇𝐶 Or = (3.6) 𝑑𝑄 𝑑𝑄 That is MR = MC (3.7) The necessary condition of profit maximisation is MC = MR. where MC denotes marginal cost and MR denotes marginal revenue. 30 University of Ghana http://ugspace.ug.edu.gh For a processor in perfect competitive market to maximise his/her profit, the marginal cost (mc) should be equated to the set market price (p). 𝑀𝐶 = 𝑃 (3.8) The second order condition should also be satisfied. This can be specified as: 𝑑2𝜋 < 0 (3.9) 𝑑𝑄2 If the first and second order conditions are satisfied the farmer is said to have maximised his/her profit. According to Sankhayan (1988), a production function with m variable inputs (X1, X2, ……, Xm) and n fixed inputs (Z1, Z2,…….,Zn) is related to output. Following Taphee et al. (2015) the generalised profit function is given as: 𝜋 = 𝑃𝑦𝑓(𝑋1, 𝑋2, … . 𝑋𝑚, 𝑍1, 𝑍2, … . 𝑍𝑛) − ∑ 𝑃𝑖𝑋1 (3.10) Where Py is the price of output and Pi is the price per unit of the ith variable input. The revenue equation is given as: 𝑇𝑅 = 𝑃𝑦𝑌 (3.11) The cost equation is given as: 𝑇𝐶 = 𝑃1𝑋1 + 𝑃2𝑋2 + 𝑃3𝑋3 + 𝑃4𝑋4 + 𝑃5𝑃5 (3.12) According to Taphee et al. (2015) the specified profit function is given as: 𝜋 = 𝑃𝑦𝑌 − (𝑃1𝑋1 + 𝑃2𝑋2 + 𝑃3𝑋3 + 𝑃4𝑋4 + 𝑃5𝑋5) (3.13) Where 𝜋 denotes Profit, PyY denotes Value of palm oil produced, P1X1 denotes Cost of labour, P2X2 denotes Cost of milling, P3X3 denotes Cost of palm nuts, P4X4 denotes Cost of water and P5X5 denotes other Cost. 31 University of Ghana http://ugspace.ug.edu.gh The profit function is then expressed as a logarithmic form of the profit function. To capture agglomeration, the profit function is expressed as a function of cluster of the processors in the study area. 3.3 Conceptual Framework of the Study The palm oil industry in Ghana as captured in figure 3.1 is an income-generating activity that provides a livelihood for resource poor women in the Eastern Ghana. Sulkava et al (2015) said that profitability of farm enterprises is very useful and as such makes it possible for the farms to stay in business in the long run and, thus, be a part of a stable food value chain. The study considers clustering as an enhancement of profitability. Porter (2000) in his study suggested that, a successful local cluster depends on four main factors namely : (1) context for firm strategy and rivalry inside the cluster, e.g., competition and collaboration put pressure on productivity; (2) demand conditions, e.g., level of sophistication and demand of consumers; (3) related and supporting industries, e.g., the supporting suppliers and ancillary industry; and (4) factor conditions, e.g., availability of infrastructure, skills and capital .Therefore when small scale enterprises cluster, they can enjoy these benefits such as shared intelligence, shared customers, competition, labour pooling, storage infrastructure. This will contribute to their technical know-how; good management system and their production capacity hence increase in income and profit. 32 University of Ghana http://ugspace.ug.edu.gh Figure 3.1 Conceptual framework: Linking clustering to profitability Source: Author’s illustration (2018) 3.4 Methods of Data Analyses The specific objectives of the study were to describe the nature of clustering of small scale palm oil processors, examine the profitability level among the clusters and examine the effect of clustering influence on profitability among small-scale palm oil processing enterprises in the Birim North District in the Eastern Region of Ghana. In the following sections the methods used to execute each specific objective have been described. 33 University of Ghana http://ugspace.ug.edu.gh 3.4.1 Describing the nature of clustering of small-scale palm oil processors in the BND Descriptive statistics such as frequency distribution tables and percentages were used to demonstrate the socio-economic characteristics of the respondents. The socio-economic characteristics included their business form, whether their businesses are registered, number of employees, years of business existence, qualification of manager and average monthly sales. The study also determined the extent to which small scale palm oil processing enterprises are clustered. Factors determining clustering such as to access common suppliers, common labour, common market, transportation, infrastructure, business information and common input materials were presented to respondents. Principal Component Analysis (PCA) was used to analyse the extent to which small-scale palm oil processors are clustered after taken the mean of the various variables. The multivariate approach used to develop the extent of clustering involved the use of Principal Component Analysis (PCA) and Cluster Analysis (Bidogeza et al., 2009; Dossa et al., 2011; Diniz et al., 2013; Nainggolan et al., 2013). The components from PCA were retained and input into K-means clustering technique which follows the approach by Ding and He (2004). Gelbard et al. (2007) guideline of PCA was used for this study. Following Filmer and Pritchett (1998), Vyass and Kumaranayake (2006) and Achia et al. (2010), PCA was applied on the dummy variables, to identify the dimensionality of the data (Jolliffe, 2002). The seven activities used in the PCA were common suppliers, common labour, common input materials, business information, infrastructure, transportation and common market. PCA is a multivariate statistical method used to decrease the number of variables into a smaller number of ‘dimensions’, with minimal loss of information. The first new variables account for as many variations in the original data as possible (Jolliffe, 2002; Manly, 2005). 34 University of Ghana http://ugspace.ug.edu.gh The new variables are linear combinations of the original variables. The suitability of the variables for PCA was checked by the Kaiser-Maier-Olkin (KMO) and the Bartlett’s sphericity tests. According to Hair et al. (2006), the variables are considered suitable if the KMO values are greater than 0.5 and Bartlett’s sphericity test is at p<0.05. In choosing the number of PCs to retain, the criterion used involved selecting the Eigenvalue that allowed for more sampling variation. The Kaiser's rule, that is Eigenvalue equal to one, would retain too few variables. Therefore, an Eigenvalue above 1 was used as the cut-off (Jolliffe, 2002). Chibanda et al. (2009) as cited in Garson (2011), recommends hierarchical clustering for dummy variables and for data sets with a sample size not less than 200. According to Kaur and Kaur (2013), the K-means algorithm performs better than the hierarchical algorithm on a large data set (i.e., greater than 100). K-means analysis was, therefore, appropriate for the sample size of the present study. Principal Component (PC) scores were used for the K-means cluster analysis as the second part of the multivariate approach to classifying the components into typologies. The PC scores are continuous solutions to the discrete cluster membership indicators for K-means cluster analysis (Ding & He, 2004). According to Jolliffe (2002), cluster analysis can be used on data which has no clear group structure. Based on the objective to examine the nature to which small scale palm oil processing enterprises are clustered, cluster analysis was used to group them. The aim of this technique was to identify and classify the respondents into a reasonable number of clusters that best explain the extent of clustering. 35 University of Ghana http://ugspace.ug.edu.gh 3.4.2 Estimating the profitability level among clusters in palm oil processing Profitability assessment in this study employed Gross Margin which requires assessing the variable cost as well as revenue for a product in a period of reference. The period of reference for this study was 2017 production year. Gross revenue (GR) accounts for all revenues earned from sale of palm oil. Palm oil processors were asked to recall the sale of palm oil within the stipulated reference period. Variable cost (VC) in this study is the cost that varies with output within the observation period. Variable cost included expenditures on labour, water, palm nut, milling and other cost. Also included in the variable cost was imputed cost of items that was not purchased but used by the palm oil enterprise from their household or received from elsewhere. This was estimated by the palm oil processor on the basis of how much he/she would have paid if it was purchased. Mathematically, total variable cost of farmer i is given as: VCi  Labcos ti Water cos ti  PalmNut cos ti Mill cos ti Other cos ti (3.14) Where, Labcos ti denotes cost of labour, Water cos ti denotes cost of water, PalmNut cos ti Mill cos ti denotes Milling cost and Other cos ti denotes the other cost. Gross Margin can be ascertained by the difference between total variable cost (VC) and gross revenue (GR) of production. For a given palm oil enterprise i, the Gross Margin is algebraically expressed as: NIi GRi VCi (3.15) Where NI denotes Net Income Benefit cost ratio is estimated as the ratio of the benefits to the total cost. The BCR is expressed as: 36 University of Ghana http://ugspace.ug.edu.gh n Benefits BCR  (3.16) i1 cos t BCR result of greater than one is an indication of the financial profitability of the small scale palm oil processing enterprise. 3.4.3 Determining the effects of cluster influence on profitability of small scale processors Following Taphee et al. (2015), a profit function was used to determine the effects of cost associated with palm oil production on the profit obtained. Augmented Cobb-Douglas production function was employed to determine the influence of clustering on palm oil firm profitability. The Augmented Cobb-Douglas production function was employed due to is simplicity. y  x '  (3.17) Where y = Dependent variable x = regressor  = error term The Empirical Model is specified as: ln Pr ofit  0  1Gen 2 ln Age 3 ln Edu  4 ln HHsize 5Occu  6Cred  7 ln Mill Cos t 8(PalmCos t) 9 (Cluster) 10(LabCos t) 11(Water Cos t) 12 (Other Cos t) The description, measurements and a priori expectations of the variables are presented in table 3.1 37 University of Ghana http://ugspace.ug.edu.gh Table 3. 1: Description of variable for effect of cluster influence on profitability Variable Description Measurement Expected sign Dependent Variable (Ln ) Log of Gross Margin GHS NA Independent Variables Socio-economic factors Gen Gender 1= Female +/- 0= Male LnAge Log Respondent’s Age Years - LnEdu Log of Education Years of + education Institutional factors Cred Access to credit 1= Yes + 0= No Input factors LnMillCost Log of Milling Cost GHS - LnPalmCost Log of Palm nut Cost GHS - LnLabCost Log of Labour Cost GHS - LnWaterCost Log of water Cost GHS - LnOtherCost Log of Other Cost GHS - Other factor clus_ Cluster 1= Cluster 1 + 2= Cluster 2 3= Cluster 3 Source: Author’s compilation (2018) Hypothesis Testing t-Statistic was employed to test the significance of hypothesis in relation to the different independent variables. 𝛽 𝑡 − 𝑆𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 = 𝑖 (3.18) 𝑆𝐸(𝛽𝑖) (Ho): Clustering has no effect on profitability (Ha): Clustering has effect on profitability This also applies to age of the processor, household size, occupation of processor, educational level of the processor and access to credit. Decision rule: If Tcal >Tcrit reject the null hypothesis. 38 University of Ghana http://ugspace.ug.edu.gh 3.5 Method of Data Collection Data type and sources: Data was gathered from primary and secondary sources. Primary data was collected through administration of structured questionnaire (appendix 1). The secondary data for the study were derived from Ghana Statistical Service (GSS), Statistics, Research and Information Directorate (SRID) of Ministry of Food and Agriculture (MoFA), FAO website (FAOSTAT). Instrument used: The main tool used to gather data was questionnaires. The questionnaire was constructed to solicit data and information in response to each research question. Four sections of the questionnaire were created. First section solicited socio-economic characteristics of respondents. Second section solicited data on the extent to which small scale oil processing enterprises are clustered in Birim North in the Eastern Region. Third section solicited data on the profitability levels among the clusters of small-scale palm oil processing enterprises in Birim North in the Eastern Region. The last and fourth section solicited the effect of clustering on income of small scale palm oil processing enterprises in Birim North District of Ghana. Target population: The target population was small scale palm oil processing enterprises in the Birim North District in the Eastern Region of Ghana. Its importance to the study is that the study area is one of the ten administrative regions in Ghana where variety of agro processing activities are found. Burger et al (1999) in their study in Indonesia on clustering of small agro-processing firms classified at least 20 firms in a village as clustered. Therefore, based on this literature at least twenty (20) small scale palm oil processors in New Abirem were considered as clustered. 39 University of Ghana http://ugspace.ug.edu.gh Sample size and sampling technique: The study was conducted in Birim North District (BND) in the Eastern Region of Ghana where Agriculture is the mainstay of the economy. MoFA (2010) indicated that about 74% of the labour force in the district is engaged in one form of agricultural enterprise or the other. Existence of big companies into agriculture, especially palm oil production and existence of Agricultural Research Institutions both within and outside the district offer good opportunities for agricultural production. Palm oil production is the major activity in the district. Small scale processors numbering about 120 in the district are engaged in extraction and marketing of palm and kernel oil. In choosing the respondents for the study, a two-stage simple random sampling technique was employed. In the first stage, the technique was used to select the community of palm oil processing towns in the districts. In the second stage, 200 processors within the community were interviewed using a structured questionnaire. New Abirem, was selected basically because there are many processing centres in the area. 3.6 Geographical Area of Study The Eastern Region is in southern part of Ghana and is one of the ten administrative regions. The region is bounded to the east by the Lake Volta, to the north by Brong-Ahafo region, to the west by Ashanti region, and to the south by Central region and Greater Accra region. The dominant inhabitants are the Akans. The main language spoken by natives of Eastern region is Akan and Ewe. 40 University of Ghana http://ugspace.ug.edu.gh The Akosombo dam is situated in the region and the livelihood of the Eastern region is subject by its high-capacity electricity generation. The region occupies an area of 19,323 square kilometers, representing about 8.1% of Ghana's total land space. Eastern Region is the sixth biggest region in relations to land space and falls between latitudes 60 and 70 North and longitudes of 1030’ West and 0030’ East. The region is divided into 26 districts. The region experiences the greatest rainfall in June and October. The first rainy season is from May to June while the second is from September to October. The land in the Eastern region is good for the cultivation of a variety of crops including, cola-nuts, cocoa, palm oil, citrus and food crops such as yam, cassava, cocoyam, rice, maize and vegetables. The population distribution pattern shows that 34.6 per cent of the region’s population lives in 56 urban settlements (towns with population above 5,000) while the greater percentage, 65.4 percent live in the rural communities. There are three main industrial activities in the region, namely Agriculture including Hunting, Forestry (54.9%), Wholesale and Retail trade (13.5%) and Manufacturing (9.1%). In Agriculture and related work, males constitute 57.4 per cent, compared with 52.6 per cent of females. However, females are predominant in Wholesale and Retail Trade (19.3%), compared with 7.4 per cent males. In the Manufacturing industry, female participation (9.5%) is higher than that of males (8.8%). Birim North District has the highest economically active population in Agriculture and related work for both males (77.6%) and females (73.9%) while in New Juaben, only 17.7 percent of males and 14.3 per cent of females are in that industry. 41 University of Ghana http://ugspace.ug.edu.gh Birim North District As part of the government’s-initiated decentralisation programme to promote effective and efficient management of governance and propel rapid socio-economic growth across the nation, Birim North District in the Eastern Region was curved out of the formal Birim District Council in 1987. Birim North District is privileged to be neighboured by Kwahu West Municipal at the north, Asante Akyem south and Adansi South District in the Ashanti Region flanking her at the west, hemmed at the south by Akyemansa District which is also a newly born District out of the Birim North District and to the east by Atiwa and Kwaebibirem District (GSS, 2013). The hilly areas are basically made up of lava flows and schist which in some cases rise to over 61 meters above sea level. These areas are predominated with high rainfall of about 170mm per annum. The District lies within the semi-equatorial climatic zone which is characterised by substantial amounts of precipitation. It experiences a double rainfall pattern. The amount of rainfall received in the district is between 150cm and 200cm. This high amount of rainfall and moderate temperatures has the potential of improving and sustaining agricultural activities in the district. This climatic condition supports the cultivation of cash crops such as palm oil and cocoa (MoFA, 2013) The District is a heterogeneous society with a population size of about 78,907 representing three percent of the regional population put forward by the Ghana statistical service in relation to the 2010 population and housing census and she covers a total land area of about 566.48 square kilometres. Agriculture defines a significant propelling force in driving the socio-economic development of the Birim North District. The District is dominantly characterised by rural communities and 42 University of Ghana http://ugspace.ug.edu.gh settlements. Its economy is hegemonized by the agricultural sector, which employs majority of the population. Agricultural activities can be categorised into four types in the District. These are agricultural, forestry and fishery. The major agricultural activity in the District is crop farming (97.8%), followed by tree planting (0.8%) and less than one percent of agricultural households are engaged in fish farming (0.3%) (MoFA, 2013). Figure 3.2 Map of the Study Area Source: Geography Department, University of Ghana (2018) 43 University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR RESULTS AND DISCUSSIONS 4.0 Introduction The chapter deals with the presentation and analysis of the data collected using the questionnaire administered to a sample of respondents. The chapter consists of the background analysis of the characteristics of the respondents through personal data and the analysis of all relevant data having a bearing on the research questions. The analysis was presented using tabular illustrations. 4.1 Socio-economic Characteristics of Respondents This section presents the socio-economic characteristics of respondents. The socio-economic characteristics of the palm oil processors include gender, age, level of education, and years of farming experience. Gender of the palm oil processors Table 4. 1 shows that about 96 percent of the palm oil processors were owned by males while the remaining 4 percent were females. This implies that females involved in palm oil processing far outweigh their female counterparts. This may be due to the nature of palm oil processing as a ‘kitchen’ business. The findings of Adjei-Nsiah et al (2012) is in agreement to the fact that women dominate the small-scale palm oil processing industry. Age of the palm oil processors The age of the palm oil processors in the study area ranged from 20 to 96 years with an average age of 52 years. As shown in Table 4. 1, the highest proportion of the palm oil processors are above 60 years whereas the remaining 40% are in the economically active age group. However, 44 University of Ghana http://ugspace.ug.edu.gh the lowest proportions (5.5%) of the palm oil processors are in their active age (21-30 years). It could be inferred that younger people are migrating to find jobs in the city or they are in the village but engaged in other activities other than palm oil processing. Table 4. 1: Socioeconomic characteristics of palm oil processors in BND Variable Frequency Percentage Gender Male 192 96.0 Female 8 4.0 Age 21-30 11 5.5 31-40 43 21.5 41-50 44 22 51-60 41 20.5 Above 60 61 30.5 Level of education completed No formal education 168 84 BECE 24 12 SSCE/WASSCE 6 3 Bachelor’s Degree 2 1 Years of experience 1-10 56 28 11-20 54 27 21-30 43 21.5 31-40 23 11.5 Household Size 1-5 109 54.5 6-10 82 41 11-15 9 4.5 Source: Field Survey, 2018 Educational level of the palm Oil processors The study shows that majority (84%) of the palm oil processors do not have any formal education whereas those with bachelor’s degree had least representing 1%. Palm oil processors 45 University of Ghana http://ugspace.ug.edu.gh with BECE and SSSCE or WASSCE constitute 12% and 3% respectively (Table 4. 1). The indication here implies that non-formal education are dominant in the study area hence would require adequate education which can help them to read and understand, hence properly adopt and apply new and appropriate technologies. Years of palm oil processing experience Table 4. 1 shows that a greater percentage (28%) of the palm oil processors in the study area have had between 1 and 10 years of experience in palm oil production while palm oil processors with experience between 31 and 40 years had the least percentage recording only 11.5%. The results further show that palm oil processors with experience 11-20 years and 21-30 years recorded 27% and 21.5% percent respectively. This implies that there are good numbers of new entrants in the palm oil industry and as such would require the needed knowledge, information and technology to be able to thrive the sector. Household size Results in Table 4. 1 show that more than half of the processors (54.5%) had their household size within 1-5 followed by those with 6-10 (recording 41%) with 11-15 household size representing the least (4.5%). The average household size of the palm oil processors was 5.2. This is above the national average of 4 (GSS, 2014) and Eastern regional mean household of 3.8 (GSS, 2014). 46 University of Ghana http://ugspace.ug.edu.gh 4.2 Background characteristics of the palm oil processing industry This section presents the characteristics of the palm oil processing industry in the study area. This section highlights the number of years of existence of the palm oil processing business, business registration, type of business ownership and monthly sales. Table 4.2 Background characteristics of the palm oil processing industry in the study area Enterprise characteristics Frequency Percentage Business Registration Yes 4 2 No 196 98 Business form Sole Proprietorship 194 97 Partnership 6 3 Years of Business Existence 1-5years 32 16 6-10years 18 9 11-15years 26 13 Over 15years 124 62 Monthly Sales Below GHS 15000 146 73 GHS 15000-GHS 25000 52 26 GHS 25000-GHS 35000 2 1 Source: field Survey, 2018 Business Registration Table 4.2 indicated that small scale processors which have not registered with the RGD represents 196 (98%) of the total sample size. The table shows that only 4(2%) processors of the total sample size have registered with the RGD. It is an indication that majority of the Palm oil processors situated in Birim North have “informal” business entities. 47 University of Ghana http://ugspace.ug.edu.gh Business Ownership Out of the 200 respondents, the dominant business structure among the small-scale processors were sole proprietors representing 194 (97%) processors (Table 4.2). The next dominant form of business structure is partnership representing 6 (3%) processors. This shows that most of the small-scale processors in Birim North take the form of sole proprietorship. Monthly Sales Table 4.2 also revealed that the average monthly sales for the 146 small scale processors were below GHS 15000 representing 73% of the sample size. fifty-two respondents representing 26% records sales between GHS 15000 – GHS 25000 and two respondents representing 1% records sales between GHS 25000 – GHS 35000. Years of business existence Table 4.2 show that 16% of the processors have been in palm oil processing for between 1-5 years, 18% had 6-10 years processing experience, 26% had been in the business for 11-15 years and the remaining 62% had been in palm oil processing business for over 15 years. 4.3 The Nature of Clustering of Small-Scale Palm Oil Processors in BND This section provides the results for the clustering of the small-scale palm oil processors in the study area. In examining the nature of clustering among the small-scale palm oil processors, 7 Likert questions were asked to measure the perception of the palm oil processors towards the location of their enterprise. 48 University of Ghana http://ugspace.ug.edu.gh Perception of palm oil processors’ enterprise location towards having a common supplier From Table 4.3, majority of palm oil processors representing 78% are of the view that to a low extent they share a common supplier in the cluster whereas the least proportion (3%) of the respondents expressed the view that to a very low extent they share a common supplier despite their location of their enterprise. The results further show that 11% of the respondents believe to a very high extent they share a common supplier. 4% are also of the opinion that to a high extent they share a common supplier yet 4% of the respondents are neutral. Table 4. 3 Palm oil processors perception of their enterprise location towards having common supplier Common supplier Freq. Percent Very high extent 22 11 High extent 8 4 Neutral 8 4 Low extent 156 78 Very low extent 6 3 Total 200 100 Perception of palm oil processors’ enterprise location towards having a common labour From Table 4.4, majority of the palm oil processors (59%) in the cluster believe to a very high extent they share common labour, 38% of palm oil processors are of the opinion that to a very high extent they share common labour whereas only 1% of the oil processors is of the opinion that to a low extent they share common labour. Table 4. 4 Palm oil processors perception of their enterprise location towards having common labour Common labour Freq. Percent Very high extent 118 59 High extent 76 38 Neutral 4 2 Low extent 2 1 Total 200 100 49 University of Ghana http://ugspace.ug.edu.gh Perception of palm oil processors’ enterprise location towards having a common market Table 4.5 indicated that majority (92%) of palm oil processors are of the opinion that to a high extent they share common market whereas only 2% of the palm oil processors is of the opinion that to a low extent they share common market. About 4% of the respondents were neutral. Table 4. 5 Palm oil processors perception of their enterprise location towards having common markets Common markets Freq. Percent Very high extent 2 1 High extent 184 92 Neutral 8 4 Low extent 4 2 Very low extent 2 1 Total 200 100 Perception of palm oil processors’ enterprise location towards having a common transportation From table 4.6, majority (80%) of the palm oil processors in the cluster believe to a low extent they share transportation, 2% of the respondents are also of the view that to a very low extent they share transportation, whereas 12% of the palm oil processors is of the opinion that to a high extent they share transportation. 6% of the respondents were neutral. Table 4. 6 Palm oil processors perception of their enterprise location towards Transportation Transportation Freq. Percent High extent 24 12 Neutral 12 6 Low extent 160 80 Very low extent 4 2 Total 200 100 50 University of Ghana http://ugspace.ug.edu.gh Perception of palm oil processors’ enterprise location towards having a common storage infrastructure From Table 4. 7, majority (79%) of palm oil processors is of the view that to a low extent they share common storage infrastructure in the cluster and 10% of the respondent are of the view that to a high extent they share infrastructure. Also, 8% of the palm oil processors are of the view that their enterprise share common infrastructure with other palm oil processors in the location whereas 3% of the respondents are neutral. Table 4. 7 Palm oil processors perception of their enterprise location towards common storage infrastructure Infrastructure Freq. Percent High extent 20 10 Neutral 6 3 Low extent 158 79 Very low extent 16 8 Total 200 100 Perception of palm oil processors’ enterprise location towards having a common material From Table 4. 8, Majority of 89% of the palm oil processors in the cluster are of the opinion that to a high extent they share common material, 1% of palm oil processors are also of the opinion that to a very low extent they share common materials whereas only 7% of the palm oil processors were neutral. The remaining 3% are of the view that to a low extent they share common materials with the other processors in the cluster. Table 4. 8 Palm oil processors perception of their enterprise location towards common material Common material Freq. Percent High extent 178 89 Neutral 14 7 Low extent 6 3 Very low extent 2 1 Total 200 100 51 University of Ghana http://ugspace.ug.edu.gh Perception of palm oil processors’ enterprise location towards sharing Business Information As shown in Table 4. 9, majority of the palm oil processors (87%) are of the view that to a high extent they share business information with other palm oil processors in the area. location of their enterprise. About 7% of the processors are of the opinion that to a very high extent the share business information, 3% are neutral, 1% share business information to a low extent whereas the remaining 2% are of the view that to a very low extent they share business information. Table 4. 9 Palm oil processors perception of their enterprise location towards Business Information Business information Freq. Percent Very high extent 14 7 High extent 174 87 Neutral 6 3 Low extent 2 1 Very low extent 4 2 Total 200 100 4.3.1 Principal Component Analysis of the Nature of clustering of Small-Scale Palm oil Processors The Principal Component Analysis (PCA) was used to group the items (common supplier, common labour, common market, common transportation, common storage infrastructure, common material and business information) into those that are possibly correlated. Results from Table 4.10 showed that 39.64% of the variance explained the extent to which component PC1 is clustered. About 22.58% of the variance represented the extent to which component PC2 is clustered and component PC3 accounted for 18.27% of the variance explaining the extent of clustering. The Direct Oblimin rotation, first used however, was inappropriate for this analysis because all the three (3) factors had a component correlation matrix less than an absolute value of 0.32, 52 University of Ghana http://ugspace.ug.edu.gh such the Varimax rotation was used and each of the factors had an absolute value greater than 0.32. It was observed that all the seven (7) basis of clustering were correlated with at least one other item hence suggesting a possible and reasonable factorability as presented in the correlation table in Appendix 2. In order to determine whether or not the dataset could be factored, the KMO and the Bartlett’s sphericity tests were performed. The Kaiser-Meyer-Olkin test of sampling adequacy was 0.613 which is above the recommended value of 0.6 indicating how suited the data is for factor analysis. The Bartlett’s test of sphericity Chi-Square value of 775.483 with 21 degree of freedom was significant at the 1% level suggesting that the variables are related and as such Principal Component Analysis is useful for the analysis. Results from Table 4.10 showed that the nature of clustering was based on three components (component 1, component 2 and component 3). Table 4.10 Principle component loading estimated scores for the nature of clustering Component Eigenvalue Difference Proportion Cumulative Comp1 2.7748 1.1940 0.3964 0.3964 Comp2 1.5808 0.3020 0.2258 0.6222 Comp3 1.2788 0.4987 0.1827 0.8049 Comp4 0.7801 0.4194 0.1114 0.9164 Comp5 0.3607 0.2249 0.0515 0.9679 Comp6 0.1358 0.0469 0.0194 0.9873 Comp7 0.0889 . 0.0127 1 Number of obs = 200 Number of comp = 3 Trace = 7 Rho = 0.8049 Determinant of correlation matrix=0.019 Source: Survey data (2018) 53 University of Ghana http://ugspace.ug.edu.gh The result also indicated that 39.64% of the variance is explaining the extent to which component PC1 is clustered. About 22.58% of the variance represented the extent to which component PC2 is clustered and component PC3 accounted for 18.27% of the variance explaining the nature of clustering. In order to determine whether or not they could be factored, the KMO and the Bartlett’s sphericity tests were performed. Also, as shown in Table 4.11 the Kaiser-Meyer-Olkin test of sampling adequacy was 0.613 which is above the recommended value of 0.6 indicating how suited the data is for factor analysis. Table 4. 11 KMO and Bartlett's Test Bartlett test of sphericity Chi-square 7783 Degrees of 21 p-value 0.000 Kaiser-Meyer-Olkin Measure of Sampling Adequacy KMO 0.613 Source: Survey data (2018) The Scree plot in figure 4.1 only confirmed the components. The components were based on those that had Eigen values greater than one indicating the relationship between the components. Therefore, the three dots below that is 1.28, 1.58 and 2.77 is confirming that indeed the result from the principal loadings is true and that the clusters are based on 3 components. 54 University of Ghana http://ugspace.ug.edu.gh Scree plot of eigenvalues after pca 0 2 4 6 8 Number Figure 4.1 Scree plot of Eigen values after PCA Varimax rotations consider co-efficients which are all above 0.3. This further confirms that there is common variance between the items. In line with all these indicators, overall factor analysis was considered suitable with all the seven (7) items. From Table 4.12, the factor loadings had three components with common supplier, business information and common input material having similar loadings as one component, common labour and common market also had similar loadings as another component and transportation as the third component. Thus, the small –scale processing enterprises are based on these three components. 55 Eigenvalues 0 1 2 3 University of Ghana http://ugspace.ug.edu.gh Table 4.12 Rotated components: Varimax with Kaiser Normalization Variable Comp 1 Comp 2 Comp 3 Common supplier 0.594 Transportation 0.7261 Infrastructure Business information 0.5853 Common input material 0.5518 Common labour 0.7035 Common markets 0.7018 4.4 The K - means Clustering Results To add on, the K-means clustering was further applied to the principal component scores to obtain grouping of the variables into clearly defined clusters. Table 4.13 shows PC dimensions of the extent to which palm oil processors are grouped across the three clusters. The clusters were named based on the way in which they relate the PC dimensions of the extent of clustering. Table 4.13 Participation of processors in combination of the nature across clusters Principal Component Dimensions Cluster 1 Cluster 2 Cluster 3 Common supplier 0.547241 0.599574 -3.85151 Transportation -0.28089 -0.00387 0.878281 Storage Infrastructure 0.917813 -0.82951 0.174975 Business information -0.30301 0.282696 -0.08973 Common input material 0.26628 -0.15983 -0.24148 Common labour 0.043347 -0.03488 -0.00727 Common markets -0.00557 0.005197 -0.00165 Number of observations 80 94 26 56 University of Ghana http://ugspace.ug.edu.gh Cluster 1 represents the extent to which processors were grouped based on Common Supplier, Infrastructure, Common Input Material and Common Labour. This cluster was named “Common supplier/storage infrastructure/common input material/common labour reliant”. The groups in cluster 2 represented those groups which mainly focused on Common Supplier, Business Information and Common Market. This cluster was therefore named “Common supplier/ business info/ Common market reliant”. Finally, the cluster 3 represented the group that was based on Transportation and Infrastructure. Cluster 3 was however named “Transportation /Storage infrastructure reliant”. Differences in Socioeconomic Characteristics of among the clusters Table 4.14 reports the differences in characteristics among the palm oil processing clusters. The tests for significant difference between clusters are based on characteristics of the palm oil processor and other institutional factors. As shown in table 4.14 the average age of processors in cluster 3 recorded the highest (52.5 years) compared to their counterparts in cluster 1 (51.9 years) and cluster 2 (about 52 years). The results further show that there is no statistically significant difference among the 3 clusters. The average household size among processors in cluster 1 (5.4) recorded the highest, followed by palm oil processors in cluster 2 (5.3) with cluster 3 (4.2) recorded the lowest. There was no statistically significant difference among the 3 clusters. 57 University of Ghana http://ugspace.ug.edu.gh Table 4. 14 Differences in characteristics among the 3 clusters of palm oil processing Variable Cluster 1 Cluster 2 Cluster 3 Prob>F/Chi- square Age 51.9 52.0 52.5 0.98 Household Size 5.4 5.3 4.2 0.13 Number of years of schooling 7.6 7.0 8.6 0.49 completed Gender Male (%) 2.50 3.2 11.54 Female (%) 97.50 96.8 88.46 0.11 Years of experience 22.2 22.67 24.46 0.81 Group Membership Yes (%) 70.0 63.8 52.0 No (%) 30.0 36.17 48.0 0.18 Access to credit Yes (%) 12.5 24.5 26.9 No (%) 87.5 75.5 73.1 0.09 Female palm oil processors dominate across the 3 clusters compared to their male counterparts. The results further show that there is no relationship between gender and cluster. Among processors in cluster 1, it was observed that 70% percent belong to a palm oil association. Similar pattern was observed for processors in cluster 2 (63.8% belong to palm oil processing group) and 3 (52% belong to palm oil processing group). Credit access among the palm oil processors was low for all the clusters. As shown in table 4.15, 12.5% of processors in cluster 1 had access to credit, 24.5% was recorded for cluster 2 whereas 26.9 percent was recorded for cluster 3. This agrees with Adjei-Nsiah et al (2012) that processors were faced with lack of credit. 58 University of Ghana http://ugspace.ug.edu.gh 4.5 Profitability of Palm Oil Processing Clusters in the Birim North District Gross margin was employed to estimate the net income across the clusters. The results for the gross margin are presented in Table 4.15. The results showed that oil processors in Cluster 3 recorded the highest revenue (GHS 14,634.62) per month compared to their counterparts in Cluster 1 (GHS 9837.5) and cluster 2 (GHS 9507.45) per month. The results further showed a statistically significant difference among the 3 clusters. This agrees with Olugnuju (2008) that in using profitability analysis, palm oil processors were found to be economically viable or profitable. Onoja et al (2012) results also showed that cocoa production was profitable with mean profit of 1,634,182.72 naira. Cost of production includes (labour cost, cost of milling, water cost, cost of palm nuts and other cost including transportation and firewood). The average variable cost of palm oil processors in cluster 2 (GHS 3235.43) was found to be higher than their counterparts in cluster 1 (GHS 3012.91 per month) and cluster 3 (GHS 3173.34) per month. No statistically significant difference was observed among the 3 clusters. Cost of palm nuts was found to be the highest amongst the cost components in all the clusters. It was revealed that the average cost of palm nuts among palm oil processors in clusters 1, 2 and 3 was estimated at GHS 1439.00 per month, GHS 1510.48and GHS 1492.12 respectively. However, there was no statistically significant difference among the 3 clusters. Cost of water was found to be the least amongst the cost components in the 3 clusters. The results further show that average cost of water for palm oil processors in cluster 3 (GHS 23.15) litres per month recorded the highest compared to their counterparts in clusters 1 (GHS 21.15) 59 University of Ghana http://ugspace.ug.edu.gh and 2 (GHS 23.15) litres per month. No statistically significant difference was observed across the 3 clusters. Labour cost was found to be the highest after cost of palm nuts. The results show that cost of labour among palm oil processors in cluster 2 (GHS 1347.45) wage per month was found to be the highest followed by palm oil processors in cluster 3 (GHS 1314.23) wage per month with those in cluster 1 (GHS 1216.88) wage per month recording the lowest. No statistically significant difference was found among the 3 clusters. Table 4.15 Gross margin of palm oil processors among the clusters Items Cluster 1 Cluster 2 Cluster 3 (Amount in GHS) (Amount in GHS) (Amount in GHS) F-Statistic Average Quantity 115.73 111.85 172.17 0.004 Palm oil (liters) 85.00 85.00 85.00 Total Income 9837.5 9507.45 14634.62 0.000 Less Variable Cost Cost of Labour 1216.88 1347.45 1314.23 0.433 Cost of Milling 143.25 157.02 156.92 0.221 Cost of Palm Nuts 1439.00 1510.48 1492.12 0.689 Cost of Water 21.15 22.87 23.15 0.364 Other cost 192.63 197.61 186.92 0.716 Total Variable Cost 3012.91 3235.43 3173.34 0.536 Gross Margin 6824.59 6272.02 11461.28 0.000 BCR 3.2 2.9 4.6 The results indicate that on average, net income of palm oil processors in clusters 1, 2 and 3 remained positive hence profitable for both groups. Despite the positive net profit observed among the clusters, palm oil processors in cluster 3 (GHS 11461.28) recorded the highest compared to their counterparts in clusters 1 (GHS 6824.59) and 2 (GHS 6272.02) per month. This implies that palm oil processors in cluster 3 were more profitable than their counterparts in clusters 1 and 2. The results further indicate that there is statistically significant difference in the net income among the 3 clusters. 60 University of Ghana http://ugspace.ug.edu.gh Processors in cluster 3 recorded benefit cost ratio of 3.2. This implies that for processors in cluster 3, every GHS 1 invested in producing palm oil will yield GHS 4 profit. The benefit cost ratio recorded for processors in cluster 2 was 2.6. This implies that for an average palm oil processor in cluster 2 every GHS 1 spent on palm oil production will result in GHS 2.6 profit. Benefit cost ratio of 3.6 was recorded for processors in cluster 3. This implies that for an average processor in cluster 3, every cedi spent in the production palm oil will yield GHS 3.6 profit. Comparatively, processors in cluster 3 are more profitable compared to their counterparts in cluster 1 and 2. 4.6 Effect of Cluster Influence on Profitability of Small-Scale Palm Oil Processors Augmented Cobb Douglas production function was employed to determine the effect of clustering influence on profitability among small scale palm oil processors in the study area. Robust regression was run to ensure that efficient and unbiased estimates were produced. The empirical findings of the regression are shown in table 4.16. The diagnostics of the model revealed that it was statistically significant at 1 percent significance level. The R-square of the model recorded 0.1551 which implies that about 15.5% of the variation in the dependent variable (log of profit) is explained by the independent variables. The results show that gender was positive and statistically significant at 1% significance level. The results further show that female who engage in palm oil processing have a higher propensity to increase their profit by about 0.76% compared to their male counterparts, ceteris paribus. This is consistent with the findings of Yinusa (2005) who opined that there was a significant difference in the palm oil processing machine usage of the women in these areas than that of 61 University of Ghana http://ugspace.ug.edu.gh their male counterparts. This may be ascribed to the fact that female engage in oil processing as a ‘kitchen’ business hence earn higher net profit compared to their male counterparts. More so, females specialize in oil palm processing to allocate time, skills and resources to increase processed oil palm output. Table 4.16 Effect of clustering influence on profitability Variable Coef. Std. Err. P>t Gender 0.712*** 0.213 0.001 Re_Age 0.364 0.258 0.160 Re_Edu -0.180* 0.095 0.058 Cluster Cluster 2 0.019 0.226 0.932 Cluster 3 0.783*** 0.260 0.003 Credit 0.017 0.203 0.935 Labour_cost 1.025 0.957 0.285 Milling_Cost -1.829** 0.782 0.020 PalmNuts_cost -1.206 1.154 0.297 Water_cost 0.379 0.296 0.201 Other_cost -0.088 0.323 0.786 _cons 17.093*** 4.918 0.001 Number of obs= 200 F(11, 188) = 6.68 Prob > F = 0.0000 R-squared = 0.1551 Root MSE = 1.3568 Note: * p<0.1; ** p<0.05; *** p<0.01 Contrary to expectation, education was found to be negative and statistically significant at 10% significance level. The results show that 1% increase in educational level of palm oil processors reduces the net income by 0.18%. This could be ascribed to the fact that processors were mainly of basic education and those with higher education level do not consider palm oil processing as their main occupation. This is consistent with Nyunza and Mwakaje (2012) whose results in 62 University of Ghana http://ugspace.ug.edu.gh their study indicated that education was not significant but had a positive effect on crop profitability. Cost of milling was negative and statistically significant at 5% significance level. The results indicate that 1% increase in milling cost reduces palm oil processors profit by 1.8%. This may be explained by the high use of inputs for milling resulting in inefficiencies hence lower profit. Cluster 3 was found to be positive and significant at 1% significance level. The results revealed that palm oil processors who rely on transportation and infrastructure (cluster 3) will have their profit increased by 0.71% compared to the counterparts in cluster 1. Generally, all the clusters exhibited positive relationship with profit. Cluster 3 largely rely on transportation and infrastructure as main factors driving the of palm oil processing hence significantly increase the profit from palm oil processing compared to their counterparts in cluster 1 who largely rely on common suppliers’ infrastructure, common input material and common labor. Access to infrastructure increases the oil quality and volume hence higher income. 63 University of Ghana http://ugspace.ug.edu.gh CHAPTER FIVE SUMMARY, CONCLUSION AND RECOMMENDATION 5.1 Introduction The result of the study is summarized in this chapter. The conclusion in this chapter is based on the research results. The chapter ends with policy recommendations based on the findings and conclusions of the study. 5.2 Summary of the Study The study was carried out to assess the effect of clustering on profitability among small scale palm oil processors. Multistage sampling was employed to select the respondents for the study. Birim North was purposively selected for the study due to the existence of palm oil processing in the area. 200 small scale palm oil processors were randomly selected for the study. The study sought to describe the nature of clustering among small scale palm oil processors, estimate the profitability level among the clusters and estimate the effect of clustering on small-scale palm oil processors’ profitability. Principal Component Analysis (PCA) and K-means clustering were employed to describe the nature of clustering among palm oil processors in the study area. Gross Margin and Benefit Cost Ratio were employed to estimate the profitability level among clusters. Augmented Cobb Douglas production function model was used to estimate the effect of clustering on profitability among the clusters of palm oil processing in the study area. The summary of the study are as follows:  Results from the PCA and K-means revealed 3 clusters namely (those that rely on common suppliers, business information, common input material and infrastructure 64 University of Ghana http://ugspace.ug.edu.gh (Cluster 1), those that rely on common suppliers, business information and common market (Cluster 2), and those that rely on transportation and infrastructure (Cluster 3).  The profitability results also show that palm oil processors in Cluster 3 recorded the highest profit with about GHS 11,461.28 per month followed by Cluster 1 recording GHS 6,824.59 per month and Cluster 2 having the least with GHS 6,272.02 per month. Results from the BCR also showed that Cluster 3 (4.6) recorded the highest followed by Cluster 1 (3.2) and Cluster 2 (2.6) recording the least.  Empirical results from the Augmented Cobb Douglas model showed that Cluster 3 (transportation and storage infrastructure) was positive and statistically significant. The results further showed that Cluster 2 (common suppliers, business information and common market) did not have any effect on the profitability level of the small-scale palm oil processors. 5.3 Conclusions of the Study The following conclusions were drawn from the findings of the study:  Three main clusters exist among the palm oil processors in the Birim North District namely “Cluster 1” –rely on common suppliers, business information, common input material and storage infrastructure; “Cluster 2”-rely on common suppliers, business information and common market and “Cluster 3”- rely on transportation and storage infrastructure only.  Profitability level among the 3 clusters was positive. Comparatively palm processors in Cluster 3 were found to be the most profitable compared to their counterparts in Cluster 1 and Cluster 2. 65 University of Ghana http://ugspace.ug.edu.gh  Palm oil processors who rely on transportation and infrastructure (Cluster 3) had a positive effect on profitability. On the contrary, palm oil processors who rely on common suppliers, business information and common market (Cluster 2) had no effect on profitability. The major finding of the study is that palm oil production is profitable for all clusters. However, firms under Cluster 3 did well as compared to Clusters 1 and 2. Therefore, palm oil processors under Cluster 3 relied on transportation and infrastructure rather than on common supplier, business information and common market. Palm oil processors in the Birim North district are not profitable enough or enjoying maximum profit because they lack access to good infrastructure and better transportation system that will bring them more income hence higher profit. 5.4 Recommendations of the Study In view of the major findings of the study, the study recommends that:  State policies should encourage private sector participation of palm oil production in order to assist them with the necessary resources such as infrastructure, transportation and other social amenities (electricity and portable water). This will enable processors to derive the needed benefits of being in the cluster so as to make it more attractive for new entrants.  Government should introduce tax reliefs to palm oil processors so as to attract various investors’ involvement into the sector to accelerate the net returns of palm oil processing.  The one-district, one factory project by the ruling NPP government should consider palm oil production in the Birim North district and channel resources to the district in other to make it attractive to non-governmental organisations that are willing to invest in the palm oil sector. 66 University of Ghana http://ugspace.ug.edu.gh REFERENCES Abiola, B. O. (2006). Knowledge, technology and growth: the case study of Otigba Computer Village Cluster in Nigeria. Unpublished paper, Knowledge for Development Program, World Bank. Washington, DC. April. Abor, J., & Quartey, P. (2010). Issues in SME development in Ghana and South Africa. International Research Journal of Finance and Economics, 39(6), 215-228. Achia, T. N., Wangombe, A., & Khadioli, N. (2010). A Logistic Regression Model to Identify Key Determinants of Poverty Using Demographic and Health Survey Data. European Journal of Social Sciences, 13(1). Adjei-Nsiah, S., Sakyi-Dawson, O., & Kuyper, T. W. (2012). Exploring opportunities for enhancing innovation in agriculture: The case of Palm oil production in Ghana. Journal of Agricultural Science, 4(10), 212. Aghalino, S. O. (2000). British colonial policies and the Palm oil industry in the Niger Delta region of Nigeria, 1900-1960. Ali, M. A. (2012). Industrial clusters and social networks and their impact on the performance of micro-and small-scale enterprises: evidence from the handloom sector in Ethiopia. Aremu, M. A. (2010). Small and Medium Scale Enterprises As A Means of Employment Generation and Capacity Building In Nigeria, A Paper Presented at the International Conference on Management and Enterprise Development on “Intellectuals and New Strategies for Sustainability Development of the Third World”. Held at Conference Center, University of Ibadan, Ibadan, Nigeria. October 5th-8th. Asante, E. A. (2012). The case of Ghana's president's special initiative on oil palm (PSI-oil palm) (No. 2012: 11). DIIS Working Paper. Ayakwah, A., Sepulveda, L., & Lyon, F. (2018). Competitive or cooperative relationships in clusters: A comparative study of two internationalising agro-processing clusters in Ghana. Critical Perspectives On International Business, 14(2/3), 230-251. Ayanda, A. M., & Laraba, A. S. (2011). Small and medium scale enterprises as a survival strategy for employment generation in Nigeria. Journal of Sustainable Development, 4(1- 5), 200. 67 University of Ghana http://ugspace.ug.edu.gh Ayodele, T. (2010). African case study: Palm Oil and economic development in Nigeria and Ghana; Recommendations for the World Bank’s 2010 Palm oil strategy. Initiative for Public Policy Analysis, Lagos, Nigeria, 1-13. Bergert, D. L. (2000). Management strategies of Elaeis guineensis (Palm oil) in response to localized markets in south eastern Ghana. Master of Science in Forestry, Michigan Technological University, West Africa. (Dissertation) Bidogeza, J. C., Berentsen, P. B. M., De Graaff, J., & Lansink, A. O. (2009). A typology of farm households for the Umutara Province in Rwanda. Food Security, 1(3), 321-335. Bigsten, A., & Tengstam, S. (2011). Smallholder diversification and income growth in Zambia. Journal of African Economies, 20(5), 781-822. Boekholt, P. (1997, October). The public sector at arm’s length or in charge? Towards a typology of cluster policies. In OECD Workshop on Cluster Analysis and Cluster Policies, Amsterdam, Netherlands (Vol. 9, No. 10). Burger, K., Kameo, D., & Sandee, H. (1999). Clustering of small agro-processing firms in Indonesia. The International Food and Agribusiness Management Review, 2(3-4), 289- 299. Carrere, R. (2010). Palm oil in Africa: Past, present and future scenarios. WRM series on tree plantations, 15, 111. Chibanda, M., Ortmann, G. F., & Lyne, M. C. (2009). Institutional and governance factors influencing the performance of selected smallholder agricultural co-operatives in KwaZulu-Natal. Agrekon, (3), 293-306. Corley, R. H. V., & Tinker, P. B. (2008). The oil palm. John Wiley & Sons, London. Daddieh, C. K. (1994). Contract farming and palm oil production in Côte d'Ivoire and Ghana. Living under contract: contract farming and agrarian transformation in sub-Saharan Africa., 188-215 Danyo, G. (2013). Palm oil and palm oil industry in Ghana: a brief history. International Research Journal of Plant Science, 4(6), 158-167. De-Oliveira W. E., & Evaldo-Fensterseifer, J. (2003). Use of resource-based view in industrial cluster strategic analysis. International Journal of Operations & Production Management, 23(9), 995-1009 Dickson, K. B. (2009). Historical Geography of Ghana. Cambridge University press, London. 68 University of Ghana http://ugspace.ug.edu.gh Dike, E. (1997). Structural Adjustment and Small-Scale Industrial Entrepreneurs in South- Eastern Nigeria. Discussion paper, DP 81. UNRISD, Geneva. Ding, C., & He, X. (2004, July). K-means clustering via principal component analysis. In Proceedings of the twenty-first international conference on Machine learning (p. 29). Association for Computing Machinery Diniz, F. H., Hoogstra-Klein, M. A., Kok, K., & Arts, B. (2013). Livelihood strategies in settlement projects in the Brazilian Amazon: Determining drivers and factors within the Agrarian Reform Program. Journal of Rural Studies, 32, 196-207. Dossa, L. H., Abdulkadir, A., Amadou, H., Sangare, S., & Schlecht, E. (2011). Exploring the diversity of urban and peri-urban agricultural systems in Sudano-Sahelian West Africa: An attempt towards a regional typology. Landscape and Urban Planning, 102(3), 197- 206. Elsner, W. (2000). An industrial policy agenda 2000 and beyond—experience, theory and policy. In Industrial policies after 2000 (pp. 411-486). Springer, Dordrecht. Fafchamps, M., & Hamine, S. E. (2004). Firm productivity, wages, and agglomeration externalities. FAO. (2004a). Round Table: Financing for Agricultural and Rural Development. Prepared for the 28th FAO Regional Conference for Latin America and the Caribbean. Guatemala City (Guatemala), 26–30 April 2004. Rome. FAO. (2004b). Proceedings of the regional workshop on forests for poverty reduction: can community forestry make money? Sim, H.C. (ed.); Appanah, S. (ed.); Lu, W.M. (ed.). RAP Publication-2004/04. Regional Office for Asia and the Pacific. Bangkok, Thailand. Felzensztein, C., & Gimmon, E. (2009). Social networks and marketing cooperation in entrepreneurial clusters: An international comparative study. Journal of International Entrepreneurship, 7(4), 281. Filmer, D., & Pritchett, L. (1998). Estimating Wealth Effects without Expenditure Data--or Tears: An Application to Educational Enrollments in States of India. Policy Research Working Papers No. 1994. Foghani, S., Mahadi, B., & Omar, R. (2017). Promoting clusters and networks for small and medium enterprises to economic development in the globalization era. SAGE Open, 7(1), 2158244017697152. 69 University of Ghana http://ugspace.ug.edu.gh Fold, N. (2008). Transnational Sourcing Practices in Ghana‘s Perennial Crop Sectors. Journal of Agrarian Change, 8(1), 94-122. Fold, N., & Whitfield, L. (2012). Developing a palm oil sector: the experiences of Malaysia and Ghana compared (No. 2012: 08). DIIS Working Paper. Foli, C. (2010). History of Palm oil in Ghana. Report commissioned by the Ghana country team of the Elites, Production and Poverty research program based at the Danish Institute for Inter- national Studies. Unpublished report, not available publicly. Garson, G. D. (2011). Cluster analysis, from Statnotes: Topics in Multivariate Analysis. 2008. www2.chass.ncsu. edu/garson/pa765. statnote. html>. Acesso em, 20. Gebreeyesus, M., & Mohnen, P. (2013). Innovation performance and embeddedness in networks: evidence from the Ethiopian footwear cluster. World Development, 41, 302- 316. Gelbard, R., Goldman, O., & Spiegler, I. (2007). Investigating diversity of clustering methods: An empirical comparison. Data & Knowledge Engineering, 63(1), 155-166. Ghana Statistical Service (2015). Palm Oil Consumption in Ghana from 2011/2012 to 2014/2015 (1000 metric tonnes) Ghana Statistical Service (2018). Ghana GDP from Agriculture 2006 -2018. Trading Economics Retrieved from TradingEconomics.com. Ghana Statistical Service (2018). Statistics for Development and Progress. Provisional 2017 Annual Gross Domestic Product. Retrieved from www.statsghana.gov.gh. Ghana Statistical Service. (2014). Ghana Living Standards Survey Round 6 (GLSS 6): Labour Force Report. Gibbs, R. & Bernat, A. (1996). Rural industry clusters raise local earnings. Rural Development Perspectives, 12 (3). Gibbs, R. M., & Bernat Jr, G. A. (1997). Rural industry clusters raise local earnings. Rural development perspectives, Washington, DC: 1984 (USA). Gunu, U. (2004). Small scale enterprises in Nigeria: Their start up, characteristics, sources of finance and importance. Ilorin Journal of Business and Social Sciences, 9(1), 36-43. Gyasi, E. A. (1992). Emergence of a new Palm oil belt in Ghana. Tijdschrift voor economische en sociale geografie, 83(1), 39-49. 70 University of Ghana http://ugspace.ug.edu.gh Gyasi, E. A. (2008). Emergence of a new Palm oil belt in Ghana. Royal Dutch Society, 83(1), 39-49. Gyasi, E. A. (2008). The Palm oil Industry and Its Implications for Rural Development. The Malaysia Journal of Tropical Geography 18. Gyasi, E. A. (2011). The Palm oil industry and its implication for the rural economy in Ghana. Unpublished report, University of Ghana. Gyasi, E. A. (2012). State expropriation of land for a plantation and its impact on peasant in Ghana. Proceedings of Commonwealth Geographical Bereau Indigenous Land Rights Workshop, Christchurch, New Zealand Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate data analysis (Vol. 6). Henry, M., & Drabenstott, M. (1996). A new micro view of the US rural economy. Economic Review-Federal Reserve Bank of Kansas City, 81, 53-70. Huddleston, P., & Tonts, M. (2007). Agricultural Development, Contract Farming and Ghana's Palm oil Industry. Geography, 266-278. Ilapi Ghana. (2017) Agriculture analysis on the 2017 budget ITC. 2006a. Bringing the Poor into the Export Process: Is Linking Small Producers and Big Exporters a Solution? Agrawal, S.; India Trade Promotion Organization. ITC Executive Forum, International Trade Centre UNCTAD/WTO. Geneva, Switzerland. ITC. 2006b. Sustainable Agriculture and Value Networks: An Opportunity for Small Growers to Export Successfully? Alvarez, G. ITC Executive Forum 2006, International Trade Centre UNCTAD/WTO. Geneva, Switzerland. ITC. 2006c. Peru: Practical Cases: CPF and Productive Chains of Structured Financial Products (SFPs) Krüger, C. and Krüger, R. ITC Executive Forum 2006, International Trade Centre UNCTAD/WTO. Geneva, Switzerland. Jolliffe, I. T. (2002). Principal component analysis, ser. Springer Ser. Statist., 2nd ed. New York: Springer. Kanyua, M. J., Ithinji, G. K., Waluse, S. K. and Wairimu, W. R. (2015). Factors Influencing Profitability of Diversified Cash Crop Farming among Smallholder Tea Farmers in Gatanga District, Kenya. , (3), 87-95. 71 University of Ghana http://ugspace.ug.edu.gh Kaur, M., & Kaur, U. (2013). Comparison between K-mean and hierarchical algorithm using query redirection. International Journal of Advanced Research in Computer Science and Software Engineering, 3(7). Keeble, D. (2000). Collective learning processes in European high-technology milieux. High- technology clusters, networking and collective learning in Europe, 199-229. Keeble, D., & Wilkinson, F. (2000). High-technology clusters. Networking and Collective Learning in Europe, Aldershot et al. Komlosy, A.; Delapina, F.; Parnreiter, C.; Stacher, I. & Zimmermann, S. (1997). Ungeregeltund Unterbezahlt: Der informelleSektor in der Weltwirtschaft, HistorischeSozialkunde, 11, Journal fürEntwicklungspolitik, Ergänzungsband, 4, Frankfurt am Man: Brandes und Apsel (u.a.). Kuah, A. T. (2002). Cluster theory and practice: Advantages for the small business locating in a vibrant cluster. Journal of Research in Marketing and Entrepreneurship, 4(3), 206-228. LaFountain, C. (2005). Where do firms locate? Testing competing models of agglomeration Lynn, M. (2002). Commerce and economic change in West Africa: The palm oil trade in the nineteenth century (93). Cambridge University Press. MAFAP (2013). Monitoring African Food and Agricultural Policies, analysis of incentives and disincentives for palm oil in Ghana. Draft Version 2013. Malaysia Palm Oil Council (MPOC). (2009). Palm oil industry in Malaysia; skills and knowledge for sustained development in Africa. Retrieved from http://www.siteresources.worldbank.org/.../Malaysian_Palm_Oil_Industry Manly, B. F. J. (2005). Multivariate Statistical Methods: A Primer. Boca Rato, FL: Chapman & Hall/CRC. Marshall, A. (1890). Principles of Economics, London: Macmillan. MASDAR. (2011). Master Plan Study on the Oil Palm Industry in Ghana: Final report. Retrieved from www.masdar.com/news/ McCormick, D. (1999). African enterprise clusters and industrialization: theory and reality. World Development, 27(9), 1531-1551. McCormick, D. (1999). Enterprise clusters in Africa: Linkages for growth and development. In Enterprise in Africa-Between poverty and growth (Vol. 132, No. 143, pp. 132-143). Practical Action Publishing in association with GSE Research. 72 University of Ghana http://ugspace.ug.edu.gh Mensah, S. (2004). A review of SME financing schemes in Ghana. A paper presented at the UNIDO Regional Workshop of Financing Small and Medium Scale Enterprises, Accra, Ghana, 15TH to 16TH March 2004. Mitullah, W. V. (1999). Lake Victoria’s Nile Perch Fish Cluster: Institutions, Politics and Joint Action, Brighton IDS Working Paper 87. MoFA. (2011). Facts and Figures about Agriculture in Ghana. Accra, Ghana: Statistics, Research and Information Directorate (SRID) of MoFA. Navickas, V., & Malakauskaite, A. (2009). The impact of clusterization on the development of small and medium‐sized enterprise (SME) sector. Journal of Business Economics and Management, 10(3), 255-259. Ntsiful, K. (2010). Out grower Palm oil plantations scheme by private companies and poverty reduction in Ghana. A PhD. Thesis Dissertation. Nyunza, G., & Mwakaje, A. G. (2012). Analysis of Round Potato Marketing in Tanzania. Ofosu-Budu, K., & Sarpong, D. (2013). Oil palm industry growth in Africa: A value chain and smallholders’ study for Ghana. Rebuilding West Africa’s food potential: Policies and market incentives for smallholder-inclusive food value chains, 211-251. Oil World. (2008). Oil World Statistics. Hamburg: ISTA Mielke GmbH. Olagunju, F. I. (2008). Economics of palm oil processing in Southwestern Nigeria. International Journal of Agricultural Economics and Rural Development, 1(2), 69-77. Onoja, A. O., Deedam, N. J., and Achike, A. I. (2012). Profitability and yield determinants in Nigeria cocoa farms: Evidence from Ondo State., (4), 172-183. Osei-Amponsah, C., Visser, L., Adjei-Nsiah, S., Struik, P. C., Sakyi-Dawson, O., & Stomph, T. J. (2012). Processing practices of small-scale palm oil producers in the Kwaebibirem District, Ghana: A diagnostic study. NJAS-Wageningen Journal of Life Sciences, 60, 49- 56. Owusu, A. S. M. (2007). Living on the fast lane on an empty stomach? Eric Sakyi Nketiah (ed) Palm oil commodification and its implications on food security in Ghana (1807-2007). In E. S. Poku, K. (2002). Small-scale palm oil processing in Africa (Vol. 148). Food & Agriculture Org. Porter, M. (1990). The Competitive Advantage of Nations, New York; the Free Press. 73 University of Ghana http://ugspace.ug.edu.gh Porter, M. (1998). “Clusters and the new Economics of Competition”, Harvard Business Review, Nov.-Dec. pp (77-90.) Porter, M. E. (2000). Location, competition, and economic development: Local clusters in a global economy. Economic development quarterly, 14(1), 15-34. Pyke, F., & Sengenberger, W. (Eds.). (1992). Industrial districts and local economic regeneration. International Labour Organisation. Rola-Rubzen, M. F., Murray-Prior, R., Batt, P., Concepcion, S., Real, R., Lamban, R., & Bacus, R. (2013). Impacts of clustering of vegetable farmers in the Philippines. In Smallholder HOPES: Proceedings of the ACIAR-PCAARRD Southern Philippines Fruits and Vegetables Program Meeting (Vol. 20, pp. XX-XX). Australian Centre for International Agricultural Research. Rosenfeld, S. A. (1997). Bringing business clusters into the mainstream of economic development. European planning studies, 5(1), 3-23. Sankhayan, P. L. (1988). Introduction to the economics of agricultural production. Prentice'Hall of India Pvt. Limited. Sarku, R. (2014). Contribution of Value Addition in Agriculture to Development: Case of the Palm oil Industry in Kwaebibirem District, Ghana (Doctoral dissertation, University of Ghana). Sarku, R. (2016) Analyses of gender roles in the Palm oil industry in Kwaebibirem District, Ghana International Journal of Humanities and Social Science Vol. 6, No. 3 Schmitz, H. (1992). 'On the clustering of small firms', in: IDS Bulletin, vol. 23, No. 3, July. Schmitz, H. (2007). 14 Regional systems and global chains. Development on the Ground: Clusters, Networks and Regions in Emerging Economies, 33, 322. Schwab, K., Porter, M. E., & Sachs, J. (2002). The global competitiveness report 2001-2002. Oxford University Press, USA. Siba, E., & Söderbom, M. (2015). Enterprise agglomeration and firm performance in sub- Saharan Africa. Handbook on Trade and Development, 169. Snodgrass, D. R., & Biggs, T. (1996). Industrialization and the small firm: Patterns and policies. International Center for Economic Growth: Harvard Institute for International Development. 74 University of Ghana http://ugspace.ug.edu.gh Soyebo, K. O., Farinde, A. J., & Adetayo, E. D. (2005). Constraints of Palm oil production in Ife central local government area of Osun State, Nigeria. Journal of Social Sciences, 10(1), 55-59. Steiner, M. &. Hartmann, C. (1998). Learning with Clusters: A case study from Upper Styria. In: Steiner, M. (ed.): Clusters and regional specialisation - On geography, Technology and networks, European Research in Regional Science, 8, pp. (211- 225). Storper, M. (1997). The regional world: territorial development in a global economy. Guilford press, New York. Sulkava, M., Sepponen, A. M., Yli-Heikkilä, M., & Latukka, A. (2015). Clustering of the self- organizing map reveals profiles of farm profitability and upscaling weights. Neurocomputing, 147, 197-206. Sunil, S. & Anupam. (2014). Agro based Clusters: A Tool for Competitiveness of Indian Agriculture in the era of Globalisation. Global Journal of Finance and Management, 6(8), pp. 713-718. Research India Publications. Swann, G., Prevezer, M., & Stout, D. (1998). The dynamics of industrial clustering: International comparisons in computing and biotechnology. Oxford University Press. Taphee, B. G., Musa, Y. H., & Vosanka, I. P. (2015). Economic Efficiency of Cocoa Production in Gashaka Local Government Area, Taraba State, Nigeria. Mediterranean Journal of Social Sciences, 6(1 S1), 570. Tayler, P., & McRae-Williams, P. (2005). Internal versus external economies of scale. Working Paper, Centre for Regional Innovation & Competitiveness- University of Ballarat, Australia. Umar, H. Y., Otitolaiye, J. O., & Ibitoye, S. J. (2014). Factors Influencing Profitability among Gum Arabic Marketers in Jalingo, Taraba State, Nigeria. Global Advanced Research Journal of Social Science, 3(1), 001-007. Vyass, Seema, and Lilani Kumaranayake. "Constructing Socioeconomic Status Indexes: How to Use Principal Component Analysis." Health Policy and Planning 21, no. 6 (2006): 459- 68. World Bank. (2010). Facts about Palm oil, World Bank Group Strategy 2010. Washington D.C.: World Bank Group. 75 University of Ghana http://ugspace.ug.edu.gh Yinusa B., (2005) Investigation Of Women's Palm Oil Processing In Some Villages Of Nigeria International Journal of Manufacturing, Material and Mechanical Engineering Research Vol.2, No.1, pp.7-11, December 2015. Zigah, D. (2015). Profitability and technical efficiency analyses of small scale palm oil processing in the Assin North and South Districts of Ghana (Doctoral dissertation, Kwame Nkrumah University of Science and Technology). 76 University of Ghana http://ugspace.ug.edu.gh APPENDIX I THE EFFECT OF CLUSTERING ON PROFITABILITY AMONG SMALL-SCALE PALM OIL PROCESSORS IN THE BIRIM NORTH DISTRICT OF GHANA QUESTIONNAIRE This questionnaire is designed to collect data to be used purely for an academic purpose. The data will help the researcher to meet part of the requirements for the award of MPhil degree from Department of Agricultural Economics and Agribusiness, University of Ghana. I wish to assure you that all responses to these questions will be strictly confidential. Thank you for your cooperation and time. Instructions: Please kindly tick in the boxes provided or write in the spaces provided your responses SECTION A: SOCIO-ECONOMIC CHARACTERISTICS 1. Sex: 1. Male 2. Female 2. Age of Palm oil processor: ………… 3. Marital Status 1. Never married 2. Consensual union 3. Separated 4. Divorced 5. Married 6. Widowed 4. (a)What is the size of your household ………… (b) i) How many adult males …… ii) Adult Females…... iii) Children < 18……… 5. Highest level of education completed: 1. None 2. Primary 3.JHS/Middle 4. Secondary/Vocational 5. Tertiary (Training college/Polytechnic/University) 6.Other(Specify) 6. School grades completed (in terms of years completed) Years 7. What is your primary occupation? 1. Clerical 2. Administrative 3. Agriculture 4. Trading 5. Palm oil processor 6. Other (specify)…………. 77 University of Ghana http://ugspace.ug.edu.gh SECTION B: GENERAL INFORMATION OF THE BUSINESS 8. Is your business registered with Registrar Department General? 1. Yes 2. No 9. Business form 10. Sole Proprietorship 2. Partnership 3. Company 4. Other………………………… 11. Number of employees ……………………….. 12. Years of business existence…………………. 13. Do you belong to any Palm Oil Association? 1. Yes 2. No SECTION C: ACCESS TO CREDIT Q14a. Circle the sources Q14b. Amount Q14c.Was Q14d. If Q14e.Interest from which loans were requested from this Palm oil Yes paid per requested in the production source (If several loans processor amount anum year and provide further were requested from successful received information for each one source, show total with from source. of all request? this source requests) 1.Yes 2. No Informal Sources 1.Friends ……………………….. ………….. ………….. 2.Relatives ……………………….. ………….. ………….. 3.Religious groups …….................................. ……............ ……............ Financial Institutions 4.Banks ……………………….. ………….. ………...... 5.Government lending Institutions ……………………….. ………….. ………...... 6.Non-Bank Financial Institution …….................................. ……............ ………...... 7.Buyers directly ……………………….. ………….. ………...... (exporter, importer, trader) Others 8.NGOs ……………………….. ………….. ………….. 9.Cooperatives ……………………….. ………….. ………….. 10.Palm Oil Associations …….................................. ………….. ………….. 78 University of Ghana http://ugspace.ug.edu.gh SECTION D: BENEFITS OF CLUSTERING TO SMALL-SCALE PALM OIL PROCESSING ENTERPRISES IN BIRIM NORTH IN THE EASTERN REGION Indicate your level of agreement or disagreement to the following benefits of clustering 15. Access to raw materials 1. Easily accessible 2. Not Accessible [ ] c) other….. 16. Access to labour 1. Highly accessible 2. Not accessible [ ] c) other…… 17. Access to business advisory service 1. Yes [ ] 2. No [ ] 18. Access to accounting and management services 1. Yes [ ] 2. No 19. Training and development services. Select all that apply 1. Technical or Technology Training 2. Professional Training and Legal Training 3. Safety Training 4. Other……………… SECTION E: NATURE OF CLUSTERING OF SMALL-SCALE PALM OIL PROCESSING ENTERPRISES IN BIRIM NORTH Indicate the extent to which your business location enables you to have access to the following Very high High Low Very low Neutral extent extent extent extent 20 Common suppliers ☐ ☐ ☐ ☐ ☐ 21 Common labour ☐ ☐ ☐ ☐ ☐ 22 Common markets ☐ ☐ ☐ ☐ ☐ 23 Transportation ☐ ☐ ☐ ☐ ☐ 24 Infrastructure ☐ ☐ ☐ ☐ ☐ 25 Common input materials ☐ ☐ ☐ ☐ ☐ 26 Business information ☐ ☐ ☐ ☐ ☐ 79 University of Ghana http://ugspace.ug.edu.gh As a result of being located in clustered vicinity, did you make any of the following changes so that your business could earn more income or be more productive? How has clustering affected the following? Highly Highly Affected Neutral Unaffected affected Unaffected 27 Access to credit ☐ ☐ ☐ ☐ ☐ 28 Access to business ☐ ☐ ☐ ☐ ☐ advisory services 29 Access to accounting and ☐ ☐ ☐ ☐ ☐ management services 30 Access to labour ☐ ☐ ☐ ☐ ☐ 31 Access of Raw materials ☐ ☐ ☐ ☐ ☐ 32 Training and ☐ ☐ ☐ ☐ ☐ development services 33 Expanded size of ☐ ☐ ☐ ☐ ☐ business 34 Added new products ☐ ☐ ☐ ☐ ☐ 35 Hired more workers ☐ ☐ ☐ ☐ ☐ 36 Reduced costs by buying ☐ ☐ ☐ ☐ ☐ in greater volume or at wholesale prices 37 Improved quality of ☐ ☐ ☐ ☐ ☐ business products 38 Developed new ☐ ☐ ☐ ☐ ☐ businesses SECTION F: COST AND RETURNS 40. What quantity of palm nuts did you purchase in the last 12 months for palm oil production?..................Kg 41. What is the quantity palm oil produced in the last 12 months……………... Litres What quantity of the palm oil produced was sold in the last 12 months………… Litres 42. In the last 12 months, how much did you spend on water for palm oil processing?....................Amount in Ghana cedis 43. How much did you spend on firewood for the production of palm oil in the last 12 months?.................... Amount in Ghana cedis 44. How much did you spend on transportation on the palm oil business.................. Amount in Ghana cedis 45. How much did you spend on Milling……………… Amount in Ghana cedis 80 University of Ghana http://ugspace.ug.edu.gh 46. How much was spent on labour for threshing the palm nuts………..Amount in Ghana cedis 47. How much was spent on labour for carting………..Amount in Ghana cedis 48. How much was spent on labour for parboiling…….. Amount in Ghana cedis SECTION A: GENERAL INFORMATION OF THE BUSINESS 1. Gender Male [ ] Female [ ] 2. Age …………………………………………. 3. Household size……………………………………… 4. Is your business registered with Registrar Department General? Yes [ ] No [] 5. Business form Sole Proprietorship [ ] Partnership [ ] Company [ ] Other………………………… 6. Number of employees………….. 7. Years of business existence …………… 8. Qualification of manager No formal Education [ ] BECE [ ] SSCE/WASSCE [ ] HND [ ] Bachelor Degree [ ] Master Degree [ ] other…………...……………… 9. Number of Years of education ………………………………. 10. Access to credit Yes [ ] No [ ] 11. Average monthly sales of business…………….GHS 12. How many litres of palm oil did you produce in the last 12 months…………. 81 University of Ghana http://ugspace.ug.edu.gh NATURE OF CLUSTERING OF SMALL-SCALE PALM OIL PROCESSING ENTERPRISES IN BIRIM NORTH DISTRICT Indicate the extent to which your business location enables you to have access to the following Very High Low Neutral Very low high extent extent [3] extent [1] extent [5] [4] [2] 12 Common suppliers ☐ ☐ ☐ ☐ ☐ 13 Common labour ☐ ☐ ☐ ☐ ☐ 14 Common markets ☐ ☐ ☐ ☐ ☐ 15 Transportation ☐ ☐ ☐ ☐ ☐ 16 Infrastructure ☐ ☐ ☐ ☐ ☐ 17 Common input materials ☐ ☐ ☐ ☐ ☐ 18 Business information ☐ ☐ ☐ ☐ ☐ 82 University of Ghana http://ugspace.ug.edu.gh APPENDIX II Correlation Matrix of the items measuring Clustering Common Common Common Common input Business supplier labour markets Transportation Infrastructure material information Common supplier 1 Common labour 0.0316 1 Common markets 0.047 0.2776 1 Transportation 0.9061 0.0636 0.088 1 Infrastructure 0.823 0.1397 0.0781 0.8214 1 Common input material 0.014 0.1614 0.0173 0.0709 0.2669 1 Business information 0.0291 -0.1142 0.0232 0.1052 0.0914 0.61 1 83 University of Ghana http://ugspace.ug.edu.gh APPENDIX III PICTURES 84 University of Ghana http://ugspace.ug.edu.gh 85 University of Ghana http://ugspace.ug.edu.gh 86 University of Ghana http://ugspace.ug.edu.gh 87