Modelling farm structural change

Background and Objective

Policy and economic environment influence agricultural structural change in many ways; the details of many the related processes are still unexplored. For example, what are the impacts of globalization, and the attempts to regulate the markets? How will farm numbers develop in the future? And how will the freed resources – e.g. arable land – be re-allocated among remaining farms? So far very few existing applications try to answer these questions. Due to data availability, these studies generally use normative or ad-hoc decision rules on farm exits. Within the DFG-research group SIAG we intend to close this gap.

Approach

We combine a model which explains the probability of a farm to exit the sector on the basis of observed developments, with prospective modelling of farm adjustments and selected factor markets (e.g. the market for agricultural land).

Data and Methods

We draw on farm-individual information from farm structural surveys for 1999, 2003 and 2007 and economic information from farm accountancy data for Germany. Our model explains farm exit probability depending on current and expected future profits and the expected development of competitors (e.g. neighbouring farms competing for land). We also take into account farm and regional characteristics which influence the strategic behaviour of farms (Margarian, 2010). The econometric exit model is iteratively coupled to the representative farm group model FARMIS. This approach enables ex-ante analyses of complex policy reforms.

Results

In the scenarios analysed so far, the projection of total agricultural production is hardly affected when explicitly taking into account the impact of changing framework conditions (e.g. the milk quote regime) on farm exits – at least in comparison to a trend-based projection of farm exits. However, a first application on dairy market reform scenarios highlights the diverging impacts these may have on the developments of the number of dairy farms of different size or regions, and their income and output (Offermann und Margarian, 2014).
In the future, a coupling to market models would enlarge the range of realistic, policy-relevant scenarios which could be analysed. In our view, it will be very interesting to analyse the net impacts of selected policies, e.g. investment aid under a quota scheme, or a full liberalisation, on farm numbers.