Previous PageTable Of Contents

Why do we persist with static equilibrium whole-farm economic models?

William Bellotti

School of Agriculture, Food & Wine, University of Adelaide, Email william.bellotti@adelaide.edu.au
Future Farm Industries CRC

Abstract

There is an urgent need for more effective whole-farm analyses to address issues of productivity, profitability and sustainability. A key requirement for such an analytical framework is capacity to accommodate climate variability and uncertainty. Mathematical models of farm management based on static equilibrium conditions fail on several criteria. A more promising approach is the integration of simulation models with relatively simple economic analysis. Close participation of non-scientists (eg. farmers) is a key requirement to ensure relevance for management tasks. Success in this venture will require deep collaboration between agronomists, ecologists and economists.

Key Words

Farm management, economic analysis, uncertainty, dynamics, complexity

Introduction

Farming is becoming more complex, dynamic and uncertain. Science has a role to play in understanding and managing this complexity and uncertainty. Indeed this is a traditional role for agronomy. But have our traditional analytical tools kept pace with the demands of modern farm management? What scientific methods and approaches are available that may be useful for farm management in a variable and uncertain world?

In this paper I highlight the importance of variability and uncertainty to the task of farm management, review limitations associated with static optimisation mathematical model approaches, draw attention to the potential of simulation modelling in economic analysis, and conclude with a plea for more collaboration between ecologists, economists and agronomists to develop more relevant and useful whole-farm analyses.

Managing under variable and uncertain conditions

An early (1814) definition of agronomy from the Oxford English Dictionary emphasises management and economy (“The management of land, rural economy, husbandry”, OED, 1989). A more recent definition of agronomy is the application of science to the management of crops and soils. It is therefore not surprising that agronomy has always had a focus on the economic production aspects of agriculture. This focus on science-based management has required agronomy to be integrated with farm economics and natural resource management. In the scientific domain of systems agronomy, a key task is to integrate and synthesise across disciplines, scales and methodologies.

A key challenge for managers of rainfed farming systems is managing risk associated with low and variable rainfall. In recent years, widespread and prolonged drought has raised the likelihood that climate change is contributing to this challenge. Understanding the implications of climate variability and uncertainty for crop and pasture management has long been a central focus for agronomy. Given this background, it would be reasonable to expect that uncertainty surrounding climate variability, and/or other sources of uncertainty and variability, would be accommodated in most analyses of farm management. Yet the reality is that many mathematical analyses of farm management persist with a static analytical framework. How can an analytical methodology designed to identify an optimum solution under static conditions have relevance to an application where present climate variability and uncertainty over future change is arguably the central concern?

Adding to the problem of variability and uncertainty is the increasing complexity of farming. No longer can we assume that the sole focus of farming is orientated towards production. If we accept the view that agriculture represents a domesticated landscape in which food production is just one of several ecosystem services (Karieva et al., 2007), the need for an analytical system to understand and manage trade-offs between services becomes acute. However, it is important that in choosing a particular theoretical modelling framework we do not lose relevance and remain focussed on the key characteristics of the farm ecosystem (eg. climate variability and future uncertainty) that are central to the task of farm management.

Agronomists are concerned with productivity and resource use efficiency; for example water use efficiency, radiation use efficiency, and nutrient balances. Often, by improving efficiency we can realise improvements in both productivity and environmental outcomes. These agronomic measures of system efficiency are subject to variability and uncertainty. Understanding how management can improve system efficiency under variable and uncertain conditions is really what we are trying to achieve with our science interventions. Research into farm management, or components of farm management such as crop-pasture rotations or livestock enterprises and stocking rates, is by necessity, conducted within a conceptual framework that largely determines the research methodology to be implemented. Given the importance of the conceptual framework to the subsequent research, it is surprising that there is not more critical debate at this fundamental level.

Problems with static equilibrium optimisation approaches

Schultz (1939) emphasised the importance of change and disequilibrium to the task of farm management. As uncertainty increases, the need for flexible management responses also increases. Under these conditions the search for mathematical constants that summarise production relationships provides little assistance to the practical task of farm management. For Schultz, the role of research was to facilitate change in farmer expectations. Once expectations are formed, planning and management becomes relatively straightforward.

Dillon (1979) reflected on the scientific discipline of Farm Management research in Australia and concluded that it had lost touch with practical farming because of “logically attractive but largely inapplicable theory”. Like Schultz before him, Dillon emphasised the management challenges created by uncertainty and dynamics, and the failure of existing mathematical models of farm management to capture these features adequately. However, Dillon did see a continuing role for farm management models at an aggregate scale for policy analysis.

McCown et al. (2006) attempted to attract wider recognition of Dillon’s criticisms of the mathematical model approach embodied in Farm Management. Following Dillon, they listed five reasons why static equilibrium models can not be applied to the practice of farm management:

  • Individual farms are unique.
  • Farm systems are dynamic.
  • Farm systems are complex.
  • Farming is conducted under conditions of uncertainty.
  • Individual farmers have different preferences.

Despite this long history of scientific criticism, mathematical models designed to identify some optimum solution under static equilibrium conditions, remain popular with farm economists for analysing farm innovation and agricultural policy (eg. Janssen and van Ittersum, 2007). Most agronomists and economists acknowledge that static equilibrium mathematical models have little relevance for practical farm management, but, they hold to the view that these models still have an important role for evaluating agricultural and natural resource management policies and research priorities. This claim deserves greater scutiny. Yes, it is true that these models can be used to evaluate policy and research priorities, but the results and conclusions of such analyses will only be valid under the assumptions of the model. For example, they can identify an optimum allocation of farm resources (land, labour, capital) under static climate conditions and certain markets. We are left with a theoretical solution for a situation that exists only in the model. Under conditions of climate variability and change, and increasing complexity, do we lose too much biological reality, and too much management relevance, by using these analytical tools?

A pragmatic hybrid of simulation modelling and spreadsheet cash-flow analysis

So what analytical approaches will be useful as we examine the trade-offs associated with alternative farm management options? In the five points listed above we see the beginnings of a specification for an analytical framework that could be useful. Agricultural production system simulation models such as APSIM (Keating et al., 2003) and GrassGro (Moore et al., 1997) overcome some of the limitations associated with static mathematical models. First (uniqueness), they can be specified for either general or specific (unique) farms or paddocks. The problem of specificity is that the transacton costs of parameterising the model for every unique paddock or farm become prohibitive. Web-based interactive provision of simulation services have been trialled with farmers to bring down transaction costs (Hunt, et al., 2006). Second (dynamics), simulation models represent the temporal dynamics associated with climate variability, some soil processes, and crop and animal growth in a dynamic, process-based approach. Typically, these simulation models operate on a daily time step, and can be run over a single season; a few years to represent, for example, a specific rotation sequence; or many years to illustrate long-term trends. Third (complexity), simulation models do not attempt to represent complex reality, rather, they focus on unifying themes, for example the soil-plant-atmosphere water continuum, so that complexity can be understood and managed. Complexity can always be increased, but this comes with increased costs for training overheads, software maintenance, and model parameterisation. Furthermore, there is no simple association between increasing complexity and utility for management practice. Currently temporal dynamics and complexity are represented relatively strongly, while spatial complexity is represented relatively weakly. Multiple-paddock versions of simulation models (eg. AusFarm, AD Moore, pers. comm.), are in the early stages of development and evaluation but hold promise for analysing crop-livestock farming systems requiring spatial awareness. Fourth (uncertainty), simulation models can assist in the management of uncertainty by allowing alternative management scenarios to be evaluated under variable, dynamic and complex environments.

Simulation models have no inherent advantage with regard to the fifth point (individual preferences), but here there is an opportunity to utilise output from simulation models as input into economic analyses that can accommodate the preferences of the individual. The challenge is to present simulation model output in a form relevant to the requirements of individual farm managers. It is this opportunity that has led to an interest in linking simulation models to relatively simple spreadsheet economic models. Three implementations of this approach are described below.

The GrassGro model simulates livestock production from grazed pastures. Pasture growth is simulated with a soil water balance, driven by daily climate data (Moore, et al., 1997). In addition to a wide range of biological outputs, a very useful feature for managers of livestock enterprises is an annual gross margin calculator based on user nominated costs and prices and biological quantities simulated by the model. The influence of climate variability on financial risk associated with different management strategies is illustrated by running the model with historical climate data over several decades. In this way, for example, the risk of additional costs of supplementary feeding associated with higher stocking rates can be understood and managed.

McCown and Parton (2006) describe a similar analytical framework where output from simulation models is used to populate gross margin spreadsheets, but they take additional steps to improve the relevance of the analysis for farm management decisions. Costs and prices are provided by farmers, local relevance is enhanced by initialising simulations with soil water and nitrate data monitored in farmer’s fields, seasonal variation is quantified with dynamic simulation, and farmer nominated management options evaluated. A useful feature for farmers is the familiar gross margin spreadsheet, but enhanced in this case by inclusion of climate variability and intrinsic representation of the riskiness of the management task.

This approach has also been applied in a smallholder crop-livestock system in Indonesia (MacLeod, et al., 2008). In this example, the APSIM simulation model provides biological quantities (eg. crop and forage yields, crop stubbles) for spreadsheet models of livestock production and household economy. The overall analytical framework provides a capacity for integrated analysis of traditional and proposed new production systems. The household livelihood component provides cash-flow, gross margins, and net income analyses. This system analysis framework allows a range of scenarios to be explored by local farmers and scientists prior to ‘proof-of-concept’ evaluation by local farmer groups.

A common feature in these examples is the close involvement of non-scientists (farmers) as partners in the research process. This participation of non-scientists is facilitated in part through the application of simulation models. Traditional discipline-based science embodied in the simulation models is made accessible to non-scientists through appropriate user-interfaces, for example through the provision of economic analyses relevant for management decisions.

Despite several decades of research and development, the theory and application of agricultural systems analysis tools is still at an early stage of development. Australian scientists have made several significant contributions, but many challenges remain. Examples where future advances are needed include:

  • Whole-farm cash-flow analyses, particularly for analysing the transition from existing to new farming systems. For example, business implications of including more perennial vegetation on grain producing farms.
  • Multi-paddock analyses, important for analysing livestock systems, crop rotation systems, and integrated crop-livestock systems.
  • Understanding and managing trade-offs between ecosystem services. How do we value traditional provisioning ecosystem services (eg. food production) along with supporting, regulating and cultural ecosystem services (eg. carbon, water, biodiversity, aesthetics) to move towards sustainable agricultural landscapes?
  • Life-cycle analysis. For example, how does biologically fixed nitrogen compare with fertiliser nitrogen when whole of life-cycle costs of fertiliser production, transport, and carbon emissions are taken into account?

These few examples illustrate the need for closer communication and collaboration between agronomists, ecologists and economists. Moving across traditional discipline boundaries is difficult and requires goodwill from all involved. Modelling frameworks have an important role in facilitating communication across traditional disciplines. The urgent need for practical solutions to real world problems will foster these collaborations.

Conclusion

As variability, complexity and uncertainty become more important and central to the task of farming and agricultural research, the value of economic models of farm management based on static equilibrium concepts is questioned. This current questioning of the value of static mathematical models follows a long history of logical criticism by eminent economists and agronomists. A promising development is the integration of agricultural simulation models with relatively simple gross margin and/or cash-flow economic spreadsheet models. This pragmatic approach combines the benefits of simulation models (dynamics, complexity, uncertainty) with the utility of spreadsheet economic models (relevance, timeliness, familiarity). There is a need for greater investment in agronomic systems research to understand and manage trade-offs between ecosystem services. This agricultural system research is needed to provide a synthesising and integrating balance to rapid developments in reductionist science and commodity driven research. Future challenges will require much deeper collaboration between farmers, agronomists, ecologists and economists.

Acknowledgement

The author would like to acknowledge the support of an OECD Fellowship from the Biological Resource Management for Sustainable Agricultural Systems Program that allowed him to undertake Study Leave with the Centre for Production Ecology and Resource Conservation, Wageningen University and Research, The Netherlands.

References

Dillon, JL (1979) An evaluation of the state of affairs in Farm Management. South African Journal of Agricultural Economics, 1, 7-13.

Hunt, J et al. (2006) Yield Prophet: An online crop simulation service. Proceedings of the 13th Australian Society of Agronomy Conference. Perth.

http://www.regional.org.au/au/asa/2006/concurrent/adoption/4645_huntj.htm

Janssen, S and van Ittersum, MK (2007) Assessing farm innovations and response to policies. Agricultural Systems, 94, 622-636.

Karieva, P Watts, S McDonald, R Boucher, T (2007) Domesticated nature: Shaping landscapes and ecosystems for human welfare. Science, 316, 1866-1869.

Keating, BA, et al., (2003) An overview of APSIM, a model designed for farming system simulation. European Journal of Agronomy, 18, 267-288.

McCown, RL, Brennan, LE, Parton, KA (2006) Learning from the historical failure of farm management models to aid management practice. Part 1. The rise and demise of theoretical models of farm economics. Australian Journal of Agricultural Research, 57, 143-156.

McCown, RL and Parton, KA (2006) Learning from the historical failure of farm management models to aid management practice. Part 2. Three system approaches. Australian Journal of Agricultural Research, 57, 157-172.

MacLeod, N, Lisson, S, McDonald, C, Corfield, J, Rahman, R, Puspadi, K (2008) Integration of smallholder crop-forage-livestock systems in South East Asia – an eastern Indonesian case study. Proceedings of XXI International Grasslands Congress, Hohhot, China, In Press.

Moore, A.D., J.R. Donnelly & M. Freer (1997) GRAZPLAN: Decision support systems for Australian grazing enterprises. III. Pasture growth and soil moisture submodels and the GrassGro DSS. Agricultural Systems, 55, 535-582.

Oxford English Dictionary (1989) Accessed online. http://dictionary.oed.com/cgi/entry/50004717?query_type=word&queryword=agronomist&first=1&max_to_show=10&single=1&sort_type=alpha

Schultz, TW (1939) Theory of the firm and farm management research. Journal of Farm Economics, 21, 570-586.

Previous PageTop Of Page