06Previous Supply Elasticity Estimates for Australian Broadacre Agriculture

Griffith, G.R., I'Anson, K., Hill, D.J. and Vere, D.T. (2001), Previous Supply Elasticity Estimates for Australian Broadacre Agriculture, Economic Research Report No. 6, NSW Agriculture, Orange.

Executive Summary

Reliable estimates of the responsiveness of the supply of and demand for agricultural products to prices and other factors are fundamental to accurate economic forecasting, valid analyses of the impacts of new production technologies or promotion campaigns, and effective policy decision making. This requirement holds true whether the estimates are used by academics, government departments such as NSW Agriculture, research institutions such as ABARE or the Beef CRC, or producer organisations such as MLA.

This paper reports a listing and review of some previous supply elasticity estimates for major Australian broadacre agricultural products. The review includes some of the early quantitative work from the 1960s (such as the survey by Gruen et al. 1967), the major programming studies of the 1970s (such as APMAA and RPM as reported in Wicks and Dillon 1978 and Longmire et al. 1979, respectively), and the mainly econometric studies of the late 1980s and 1990s (such as EMABA, Wall and Fisher 1987, and Kokic et al. 1993). However, not all of the studies that have been conducted in the area have been able to be covered in the review.

The studies reviewed vary substantially in terms of geographic coverage, sample periods, estimation method, functional form, other explanatory variables included and reliance on the underlying production theory. Data limitations restrict the majority of studies to estimates of aggregate supply elasticities, although most studies do break down the estimates into different states or agro-climatic Zones. There is a mixture of estimates by length of run. Inspection of the estimates in Tables 1, 2, 3 and 4 reveals major discrepancies. This is indicative of the lack of consensus regarding supply response in Australian agriculture.

An unresolved issue is the relative merit of the mathematical programming and econometric models. In terms of the relative outcomes of the two model types, the main observable difference is that elasticities generated from programming models are generally higher than those from econometric models. Hall, Fraser and Purtill (1988) give several reasons why such a difference can be expected. Programming models permit a higher level of disaggregation, which has served to illustrate variations in supply response by region and by farm type that would be hidden by an aggregate model. However, Kokic et al. (1993) have attempted to bring this disaggregation into an econometric model, using AAGIS information at the farm level to provide a highly detailed cross-sectional picture of broadacre agriculture.

There is consensus that when the price of a product rises, the response in supply takes two forms. The first is the expansion effect or the net increase in output of one or more products, and second is the transformation effect which reflects the change in the mix of products along the production frontier, resulting from the greater relative profitability of the product whose price has risen. Generally, as a measure of the expansion effect, own-price supply elasticities for the four products covered in this review are inelastic, although some of the programming estimates exceed unity in the medium to long run. The wheat estimates tend to be larger than the livestock estimates, as there is more flexibility to alter cropping acreages than livestock numbers, especially in the short to medium run. Similarly, the estimates for the Wheat-Sheep Zone tend also to be larger than either the High Rainfall or Pastoral Zones, as transformation possibilities are greater in the former region. There is little agreement over the values for cross-price elasticities of supply, because there are a variety of assumptions used to restrict the values or the signs of the transformation effects. Vincent, Dixon and Powell (1980) assumed the expansion effect to be positive, and the transformation effect to be negative. Wall and Fisher (1988) used similar assumptions. The ORANI model (Adams 1987) in its treatment of inputs as non-specific to outputs, expresses the jointness of production in terms of production systems or composite products. Cross-price elasticities for these composite products have been constrained to be negative, but no such constraint has been made for individual products, and under certain circumstances, the crossprice elasticity for transformation may also be positive. In the econometric results, some crossprice elasticities are unconstrained while others have been constrained to be positive. Generally however, cross-price elasticities are also inelastic and mostly negative, although sheepmeat and wool are often estimated to have a positive cross-price elasticity as they are joint products.