Identifying the typology of tea smallholding development societies in Matara district, Sri Lanka

Abstract

Tea Smallholding Development Societies (TSDS) are farmer organizations in the tea smallholding sub-sector in Sri Lanka, established through a Parliamentary Act in 1991. They are managed by an elected committee and set up to assist members on extension, inputs, marketing, welfare and credit facilities. Evidence shows that the tea smallholding sub-sector is facing a crisis, which implies that there is a shortfall in TSDSs’ performance. This study attempts to identify and classify the TSDSs in Matara district through multifunctional activities. The efficiency of TSDS was assessed based on its level of intervention in twelve activities. A sample of 25 TSDSs was selected from Matara district from three strata identified through preliminary studies. Samples were drawn randomly within the strata. Data were collected using structured interviews with a standard questionnaire. TSDSs were assessed using the 0-10 score scale based on the level of engagement in identified activities. The majority of TSDS (72%) in the district show the lowest level of efficacy (≤4 score) in terms of the execution of multifunctional services. The basic functions such as identification of the needs of members and administrative activities are not appropriately implemented by most of the TSDS in the region. In reviewing the outcomes of General Meetings, there appears to be a lack of policy dialogue within organizations. Production support packages (inputs, resources, credit, and extension) are poorly implemented by the majority of the TSDSs and only 4-16% of TSDSs received satisfactory scores (≥5). Market-oriented activities are at a very low level (≤2) and can be seen only in two TSDSs. Although welfare programmes are motivational, only 28% have implemented such to enhance the livelihoods of the members. Sign test revealed that the sample did not perform satisfactorily on any of the reviewed tasks. A cluster analysis indicated three clusters. Relatively highest centroid values for variables were observed in the first cluster (C1), which has two TSDSs. The most poorly performing cluster (C2) has 64% observations. It is noteworthy that the centroid values of only two variables were exceeded five, even in the highest performing cluster. Further exploration of inter-cluster variability would help to understand the factors that make them effective.