The objectives of this study are threefold. First, to review some of the methodological aspects of farm growth research, with particular emphasis on systems simulation. Second, to develop a model which is representative of individual pastoral firms in the Land Development (Fitzroy Basin) Scheme, and to use this model to identify optimal growth strategies for land selectors. Third, to appraise the suitability of the simulation technique for farm growth research and planning and the scope for wider application in the light of the results and experience obtained in the empirical study.
The methodological review reveals that systems simulation has advantages over alternative growth research techniques such as mathematical programming for the modelling of complex bio-economic systems, particularly with regard to incorporation of uncertainty, multiple goals and non-linear relationships. However, lack of optimizing ability and high cost have been viewed as deficiencies of the simulation approach. The first of these limitations has been at least partially overcome in this study, and a number of steps have been taken to minimize cost.
A model of the farm-firm obtained under ballot in Area 3 of the above project (commonly known as the Brigalow Scheme) is presented and discussed. This model has been designed to simulate farm performance under various growth strategies, defined in terms of annual rates of land development, regrowth control and breeder purchases in the first five years of selection, during which time external finance of the order of $150,000 is utilized by settlers. The farm growth model is stochastic in that random sequences of weekly rainfall and annual beef prices and cost levels are generated for use in growth simulation experiments. Particular attention is paid to correlation between these variables and to autocorrelation within their levels over time, and sampling procedures have been devised to incorporate these correlations in generated sequences. Available soil moisture predicted from rainfall levels is monitored to determine weekly pasture conditions and consequent livestock performance. A detailed accounting model records annual incomes, growth costs and operating expenses of the farm-firm. Firms are forced into simular bankruptcy should debt servicing arrears exceed a critical level.
Survey evidence reveals that settlers have goals of growth in net worth, security (low failure risk) and annual consumption targets. These goals are combined into a lexicographic multidimensional objective function. The farm growth model is, in effect, a procedure for predicting the level of this objective function for a given growth strategy. It can therefore be used to experiment with different management alternatives with a view to identifying the optimal growth path. This is achieved by a systematic search over the policy space or spectrum of feasible alternatives. The relatively efficient and highly robust method of conjugate directions has been used to identify conditionally optimal growth strategies for various initial resource situations.
While the scope for testing the predictive ability of the overall model was severely limited, validation efforts confirmed the reliability of certain submodels. Sensitivity analysis reveals that conditional optima vary with some parameters (including those governing beef prices, calving rates and income tax payments) but are insensitive to others (regrowth rate and extent and depreciation allowances for tax purposes).
Certain inefficiencies with regard to resource allocation in the Brigalow Scheme are indicated by the study. Rates of land development appear to have been unnecessarily rapid in a number of cases, perhaps due to the limited period over which development finance is made available to settlers. Inadequate initial settler's funds and scarcity of livestock finance have led to purchase of an insufficient number of foundation stock on some properties. While simulation experiments suggest that chemical control of a substantial area of scrub regrowth during the five-year period is warranted, little regrowth control has been carried out by settlers as yet.
Experience in the study confirms that systems simulation is a powerful technique for farm planning, allowing great flexibility for the modelling of complex systems. In the problem studied, this flexibility was particularly important with regard to the treatment of weather, price and cost uncertainty, estimation of failure probabilities and calculation of income tax payments. The approach was found to be expensive in terms of both model construction and implementation on the computer. Model construction through a modular approach (including use of relevant existing submodels) and development or adaption of efficient optimization procedures are proposed as methods to reduce these costs.