An empirical application of the linear programming model for agricultural land use planning: Through the valuation of negative externalities caused by farmland abandonment in marginal areas

The purpose of this paper is to introduce a method for agricultural land use planning in marginal areas. Ever since farmland abandonment became a serious problem in marginal areas, several authors have pointed out the affects of negative externalities of farm abandonment on the adjacent farmland, such as increased insect damages, weed growth, maintenance costs of common facilities and so on. However, little has been shown about a method for detailed land use planning to take into account negative externalities estimated in monetary terms. It is generally difficult to handle a regional production structure directly with empirical model approaches such as regression models assigning land use data. On the other hand, normative models, for example linear programming models, have sometimes a problem of reliability since statistical tests are not applicable. One proposal for this problem is an empirical application of a linear programming model. We construct a linear programming model that can deal with unknown values of negative externalities caused by adjacent abandoned farmland. 198 plots of rice fields and 115 plots of waste farmland are chosen as a data set from the studied area. For the first step, the linear programming model is optimized in order to calibrate the unknown unit cost of externalities. Here, the observed land use pattern is taken as a hit ratio criterion. As a result, the unit cost of negative externalities caused by farm abandonment is calculated as about 6, 300 yen/10 a. The result is able to reproduce the observed land use pattern at a hit ratio of about 76%. As the next step, a linear programming model assigning the calculated unit costs is applied to the present land use pattern. Then, a preferable future land use plan is produced. The farmland plots in the suggested land use plan are not scattered spatially, to reflect the affect of externalities. This method assumed that the observed land use patterns are derived from the result of farmers ' income maximization behavior. Accordingly, it cannot be applicable to farmlands on the urban fringe where land owners are likely to keep their lands for future development.