 Adjustment Costs of Agri-Environmental Policy Switchings: A Multi-Agent ApproachAlfons Balmann, Kathrin Happe, Konrad Kellermann, Anne Kleingarn
No 148, Computing in Economics and Finance 2001 from Society for Computational Economics Abstract: Normative agricultural policy models are traditionally either based on representative farms or on farm samples. A well-known problem of representative farm models is the aggregation error. On the other hand, farm sample models usually do not consider farms' interactions and thus, the farms' behavior may not be consistent on the aggregate level. However, the availability of powerful computers as well as the development of new modeling techniques offers new opportunities for farm sample approaches. Particularly, so called multi-agent models allow to simulate agricultural regions "from the bottom up" by considering a multitude of individually behaving farms (agents) that interact on product and factor markets, such as the land market. Purpose of this paper is to present such a normative multi-agent modeling approach and to apply it to a comparative dynamic analysis of agricultural policies that aim at reducing the environmental problems of the high animal density to be found in certain regions in Germany. The model considers a number of initially ~ 800 farms which are located at different points on a chessboard-like spatial grid representing a region of ~ 25000 hectares (ha). The cells of the grid represent land plots of 2.5 ha each on which agricultural production can take place. Farms compete for renting the land in repeated iterative auctions during which each farm bids according to its marginal land productivity and its distance to the next available plot. Apart from renting and letting land, farms can engage in different agricultural production activities (e.g. dairy, cattle, hogs, sows, poultry, arable farming, pasture land) and they can invest in different assets (differently sized buildings for various activities, machinery of different sizes). In addition to the different production and investment activities, the farms can use their labor and capital for off-farm employment as well as to hire additional labor and to make debts. Moreover, farms can give up farming and new farms can be founded. Each of the farms can be understood as an agent that acts autonomously in order to maximize the individual household income in response to expected prices, subsidies, and the availability of land. Production, rental decisions, and investment decisions are based on mixed integer programming. The model is applied to the region of "Hohenlohe" which is located in the southwestern German federal state of Baden-Wuerttemberg. The region is dominated by family farms and can be characterized by its specialization in piglet, hog, and poultry production. To grasp the particular characteristics of the region, each of the model farms is initialized with data which is derived from real farms of the region. The respective farms participate in the German Farm Accounting Data Network (FADN). On the aggregate level the modeled region is initialized such that it fits the farm structure and the production structure of the region Hohenlohe. Simulations are carried out for basic nutrient related environmental policies. National legislation, which is binding for all farms, limits the application of animal manure to farmland on a per hectare basis. However, in regions such as Hohenlohe, many farms with intensive animal production may exceed the limit. They either have to dispose excess manure by paying other farmers to take it or by renting additional land. On the state level farmers have the option to participate in a voluntary agri-environmental program in which they are financially compensated for introducing selected, more environmentally friendly production methods (e.g. a stronger coupling of plant and livestock production, introduction of extensive production methods). Switching to and increasing participation in voluntary measures is simulated. Results are interpreted with respect to environmental indicators, structural change, production structure, rental prices, efficiency, and farm incomes. Keywords: agriculture; policy analysis; agent-based systems (search for similar items in EconPapers) There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it. Related works: Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text Persistent link: https://EconPapers.repec.org/RePEc:sce:scecf1:148 Access Statistics for this paper More papers in Computing in Economics and Finance 2001 from Society for Computational Economics Contact information at EDIRC. |
Is your work missing from RePEc? Here is how to contribute.
Questions or problems? Check the EconPapers FAQ or send mail to .
EconPapers is hosted by the
School of Business at Örebro University.