prilagođeno pretraživanje po punom tekstu

ŠUMARSKI LIST 3-4/2017 str. 46     <-- 46 -->        PDF

statistical regions of the country, the dominance of certain categories of NWFPs was recognised and hence the aim of the research was clustering of the analysed companies based on selected characteristics relevant for the research. The purpose of the research was to identify clusters, establish similarities within the defined clusters (Ferligoj 2003), and diagnose the differences among them. The subjects of the research were the companies operating in the sector of NWFPs within four statistical regions of Serbia.
According to the research of Innovation Entrepreneurship in Forestry in Central Europe (Rametsteiner et al. 2005), the forestry innovation system is active in the fields of technological and organisational innovations, and in the diffusion of specific pre-selected innovations. Typical areas of activity are mechanisation of forest work and the forming of forest owners’ co-operations. With the exception of some selected topics – such as bio-energy or forest education – product and service innovations are rather disregarded. Specific support aimed at the development of new products and service innovations is practically missing (Rimmler et al. 2011; Posavec et al. 2011). Most of the non-wood products market is weak or underdeveloped, and its use is frequently unregulated due to which direct financial benefits for the owner cannot be exploited (Mavsar et al. 2008). The entire value chain from the ensurance of non-wood forest products in the forest, to the exploitation and recognition by the forest owners and users (Merlo and Paveri 1997) and improvements in forest management is considerably demanding (Pettenella et al. 2006; Posavec et al. 2010). It is influenced by both internal factors, such as ecosystem capacity and external factors, defined by interests and the requirements of the users and forest managers (Vuletić i et al. 2009). They are both influenced by polical and legal framework in the country.
In this study, the quantitative research approach was applied (Dul and Hak 2008; Lamnek 2005). Way of data collection was used as basic research technique (Lorenzini 2014). The questionnaire was intended for representatives of SMEs (i.e. owners, directors of enterprises, heads of the commercial sector, etc.). The questionnaire consisted of 46 open and closed-ended questions, divided in accordance with the analysed market elements such as: product, price and promotion. Basic database was obtained of SMEs. According to this database, 146 SMEs were registered in 2014 in Serbia and 119 SMEs were involved in purchasing, processing and sale of NWFPs which represented our sample[1]. The share of surveyed SMEs was 81.5%. In order to make a comparability of SMEs, primarily related to the socio-economic and market aspects of NWFPs, SMEs were divided by statistical regions – Belgrade, Šumadija and Western Serbia, Southern and Eastern Serbia and Vojvodina. Data collection was undertaken in the period from 2010-2015.
Since the aim of this study was to develop a typology of SMEs involved in purchasing, processing and sale of NWFPs in Serbia, the answers to questions in Table 1 were used as the basis of the typology. Due to the mixture of continuous and categorical variables, the survey data was analysed using a SPSS two-step cluster analysis as described by Norušis (2008) and Shih et al. (2010). The first step of the procedure entailed forming preclusters. The procedure first assigned cases into small sub-clusters, then scanned the data and, based on a distance measure, the SPSS software algorithms determined whether the current case had to be merged with a previously formed precluster or a new precluster needed to be started. When preclustering was completed, all cases in the same preclusters were treated as a single entity. During the second step, the preclusters were clustered by using a hierarchical clustering algorithm.
The clustering algorithm was based on a log-likelihood distance measure because both continuous and categorical variables were used in the analysis. The distance between two clusters depended on the decrease in the log-likelihood when they were combined into a single cluster (Shih et al. 2010). The identified clusters were described concerning the region, the core products, export markets, the conditions under which the manufacturing process was performed, problems in business and payment terms using cross-tabulation. Finally, the validity of the independent variables to predict cluster membership was assessed using overall model quality (measure of cohesion and separation) and discriminant analysis described by Hair et al. (2006). All statistical analyses were carried out using SPSS 20 software (Corp. 2011).
Moreover, indexes of dynamics were used (Stojković 2001) to show the relative movement of the purchase and placement of NWFPs in time. For this purpose, the average annual growth rate was used (Bartlett 1993; Liargovas and Skandalis 2008) for certain categories of NWFPs. These growth rates were determined for the purchase of NWFPs, and for the placement on the domestic and foreign markets. The structure of dominant NWFPs category was presented for each cluster formed from the point of purchase, and placement on domestic and foreign markets. The questionnaire included the features concerning the business activities of the company related to the purchase, the placement on the domestic market and the export of NWFPs. The part about products included a series of questions about the purchase of raw materials, production, placement, plans for capacity expansion and assortment. The part related to the