| dc.description.abstracteng | Social networks can serve as an instrument to link agents with each other over space and time. The emerging pattern can display the flow of information, the diffusion of technology or the trade flow of products between all kinds of agents. Networks are very flexible constructs, which are defined by their structure. Social networks enable us to overcome close spatial proximity. Thus, neighbouring agents in a network do not need to be located in a limit area.
In the first part of the thesis, the relevance of social networks in the adoption of technology is examined. As information about availability and suitability of innovation is one of the critical factors in the diffusion and adoption process of agricultural technology, its transmission is of crucial importance. Other farmers often mention farming neighbours as the most significant origin of knowledge. However, various sources, like extensions services, technology providing companies, and downstream and upstream industries, are significant stakeholders. Thus, we construct five different network structures to analyse their impact and relevance in the process of diffusion and adoption of automatic milking systems (AMS) in Germany. These five networks are the neighbourhood network, the sales structure network, the dairy factory network, the public advisory network, and the organic farming association network. A spatiotemporal endemic-epidemic model for infectious disease spread is applied to capture the dynamic contagious process of diffusion. Different outcomes occur for the networks. The neighbourhood network provides the best fit, as well as the most substantial impact on the diffusion. The public advisory network, in contrast, performs poorly and influences less than 1%. Still, the diffusion of AMS is driven by exogenous and autoregressive effects.
In the second part of the thesis, the introduced networks are tested while controlling for various factors determining the adoption of AMS in Germany. AMS are often described as a substitute for labour, because of increasing labour costs and simultaneously decreasing availability of labour in the agricultural sector. The results confirm the influence of the costs of salaries and the amount of hired and family labour on the farm on the adoption of technology. However, the network effect remains strong. Nevertheless, the autoregressive effect, which can be defined as the agglomeration effect, fosters the diffusion of technology more than the network effect. Agglomeration effects can explain the diffusion of technology in a limited spatial area (here administrative districts in Germany) due to accumulation of farms, regional spillovers, and a pooled labour market. Network effects, in contrary, are crucial to spread a technology over more considerable distances, to ensure diffusion in the whole country. Both effects together explain 50% of the diffusion of AMS in Germany between 1997 and 2013.
The third part of the thesis explores the construction of trade networks in oil palm and rubber trade in Jambi, Indonesia. In the local value chain of both products, intermediaries are positioned between smallholder producers and processors. Social networks are suitable to display the marketing strategies of those intermediaries choosing the direct or indirect path to distribute the crops. We visualise trading networks for rubber and oil palm trade using a cross-sectional dataset and a panel dataset from three survey rounds in 2012, 2015, and 2018. Rubber trading networks diminishing in size because the number of intermediaries is reducing. The oil palm trade develops in the opposite direction. Applying an exponential random graph (p*) model, we analyse factors influencing the network patten in Jambi. The average product price and the provision of loan by the buyer are the primary aspects for forming trade connections in both samples. In contrast, the trading experience and the total traded quantity do not play a significant role. However, the panel sample exhibits disparities between rubber and oil palm and between the considered years. Previous existing links in the trading network do only affect the link formation in the rubber trading network.
By considering the social environment, we are adding more layers to the economic activities of individuals. Networks can help to refine the picture we have about complex processes in agricultural economics and enable us to a better understanding of various challenges and constraints farmers are exhibiting. This thesis highlights two different applications for networks in modern agricultural systems to provide a deeper understanding of economic decision making. | de |