Data input, results preparation and data output

produced is reduced by the quantity of piglets used for fattening pig produc- tion.42

- Milk quota: Since the year 2000, prices for milk quota in Germany have been determined in quota auctions. Implementing such an auction would be a complex matter (comparable to the land auction). Regarding quota, AgriPoliS therefore implements a highly simplified quota market in that it reflects only the results of quota auctions. In principle, farms can buy and sell quota indefinitely. But, to keep milk production within realistic limits, the price of quota is related to a regional reference quota.43 If milk production is above (below) the regional reference level plus a 10% tolerance, the quota price rises (falls) by a given percentage. The quota market as implemented in the model resembles a quota leasing market. To prevent quota from leaving the region, the marginal revenue of selling quota is less than the marginal revenue of buying additional quota.

- Manure trading: Regarding manure trading, farm agents generally pay to dis-

pose of excess manure, on the one hand. On the other, farm agents receive payments for taking excess manure up to a given limit. Manure trading is not limited to the region. That is why in the simulation there may be more farms taking up manure than farms disposing of manure and vice versa. Similar to the market for milk quota, the price of disposing manure rises the more ex- cess manure is offered.

3.7 Data input, results preparation and data output

AgriPoliS has an interface to a spreadsheet file that includes data on the regional agricultural structure to be studied to initialise the model. The file contains data on individual farm agents (family labour, machinery, buildings, production facilities, land, production quota, liquid assets, and borrowed capital) as well as regional data (number of farms, farm types, total land). Figure 3-12 illustrates the procedure of reading data into AgriPoliS in a schematic way.

42

At the current development stage, there is no interdependence between the price of piglets and the gross margin of pig fattening.

43

The regional reference quota is calculated as the total number of dairy cows in the region to be modelled times the average milk yield in that region. A tolerance range of ±10% around the regional reference quota is assumed, so that it does not function as the exact threshold value for price changes.

Figure 3-12: Schematic representation of AgriPoliS input and output Input data Aggregate data for region Individual farm data Accounting data Regional statistics Data from data

collections

Input

AgriPoliS

Output

Source: Own figure.

On the input side, data – broadly speaking - input consists of farm accountancy data, regional statistics, and stylised data on technical coefficients, prices and costs. On the output side, AgriPoliS compiles aggregate data at the sector level (class SectorResults), on the one hand, and individual farm data, on the other hand. More specifically, data output at sector level and at farm level (class

DataOutput) include data listed in appendix A-2. Based on these indicators it is possible to draw conclusions with respect to production, economic perform- ance of farms, production intensity, income distribution, and farm structure.

Part II

Applying and testing AgriPoliS

4 Adapting AgriPoliS to the region Hohenlohe

4.1 Introduction

The previous chapter introduced the agent-based model AgriPoliS, which was designed to simulate the structural development of regional agricultural struc- tures. The purpose of this chapter is twofold: first, it presents a methodology for coupling AgriPoliS with data of an existing regional agricultural structure. Sec- ond, this chapter presents an adaptation of AgriPoliS to the agricultural structure of the region Hohenlohe in southwest Germany. Adapting AgriPoliS to Hohenlohe (as well as to any other region) requires the representation of key regional indicators such as the number of farms, the specific farm size distribu-

tion, farm specialisation, income sources, and production in a reference year.44

Moreover, farm agents need to be initialised based on real farm-data, for exam- ple, on production activities, capital endowment, farm specialisation, labour en- dowment. The adaptation and the model calibration focus primarily on matching the starting conditions of AgriPoliS with Hohenlohe's structure in the financial

year 2000/2001.45 The political framework conditions are given by Agenda

2000.

The adaptation of the starting conditions is done in two steps. The first step is to represent the structure of Hohenlohe based on a number of typical farms. The second step is to represent the internal organisation of each of these typical farms, that is to say, their specialisation, main production activities, asset and

44 _{The specific methodology to couple AgriPoliS with real data that is presented in this}

chapter was originally developed and tried out for the purpose of this study. It involves the replication and representation of a single reference year. The original idea to replicate a historical reference period to identify the model's parameters (as intended in the project proposal) was not followed because it proved to be very demanding (see also the discus- sion in section 4.9). The back-casting of previous development will be the subject of fur- ther research.

45 _{Calibration is the simulation of a model with different parameter values such that simu-}

capital endowments. As this chapter builds upon work by KLEINGARN (2002)

and SAHRBACHER (2003), a detailed description of the calibration procedure,

selection of data, and discussion of parameter values can be found in these refer- ences. The following summarises the most important results of these studies.