Economics of Decision Making in Pest Management

Lefter Daku
Doctoral Candidate
Department of Agricultural and Applied Economics
Virginia Tech

April 2001


Following widespread concerns about the adverse effects of pesticide use including pesticide resistance, pest resurgence, secondary pest outbreaks, effects on non-target organisms (natural enemies), and pesticide pollution since the early 60�s, it became clear that spraying by calendar was not the appropriate approach to pest control. The fundamental questions addressed were: how many insects cause how much damage, and are the damage levels all significant? Experts from a number of agricultural disciplines came to realize that a decision rule or threshold should answer such questions and that pest control ought to be viewed as a decision making process rather than as a list of individual practices for several reasons.

First, decision-making in pest management, like other economic problems in agriculture, involves allocating scarce resources to meet food demand of a growing population. In this process, agricultural producers have to make choices regarding the use of several inputs including labor, insecticides, herbicides, fungicides, and consulting expenses related to the level and intensity of pest infestation and the timing of treatment.

Second, pest management at farm level is not only related to the choice of pest control practices, but also to the optimal level of pest control by a particular practice or set of practices (Norton & Mullen, 1994).

Third, decision making process for pest control takes place in many levels at the farm and beyond including farmers, managers, sprayers, pest control advisors, researchers, government representatives involved in regulation of pesticide use, chemical industry personnel, pesticide dealers, and so forth. These various layers of decision making affect in one way or another the whole strategy of pest control in a given crop, region or country as well as the set of approaches and methods that are chosen to implement pest control programs.

Fourth, the complexity of interaction between pest populations and wider environment for a given crop and region necessitates adoption of a systems approach to managing crop pests. Research has shown that depending upon the pest complex and agro-climatic conditions, pest control programs may differ dramatically for the same crop in different regions and microzones (Mengech et al., 1995). Recognizing this complexity, Pedigo (1996) notes that bioeconomics, the study of the relationships between pest numbers, host responses to injury, and resultant economic losses, is the basis of assessment and decision-making in pest management.

While biological scientists have a primarily role in doing research on adverse effects of pesticide use on the agro-ecosystem, there is also a growing role for agricultural economists to be involved in pest management research and decision making. As Davidson and Norgaard put it (1973, p. 2) the economists� contribution in the design and development of integrated pest management strategies as well as their implications to the farmer, the agricultural sector, and society stems from the fact that: (i) the goals of pest management are largely economic, (ii) as a science of resource allocation, economics can help identify optimal quantities and combinations of pest management inputs, and (iii) economists� growing preoccupation with an institutional view of economic phenomena and processes can lead to a better understanding of incentive structure underlying farmers� behavior and the effects on these incentives of alternative institutional arrangements for speeding up adoption of integrated pest management practices.

In reviewing the major contributions on the subject, the following models are discussed and cited extensively in the literature as being the most prominent models related to the economics of decision making in pest management (Headley, 1972; Stern, 1973; Mumford and Norton, 1984; Norton and Mumford, 1993; Pedigo and Higley, 1997; Roling and van de Fliert, 1994; Norton et al., 1999): (i) the economic threshold model, (ii) the economic optimization model, (iii) the decision theory model, (iv) the behavioral decision model, and (v) the participatory model.

The Economic Threshold Model

The development of economic threshold concept and its refinements is a major contribution of entomologists. In the 60�s, researchers came up with the idea of tolerating some certain level of pest damage, reasoning that most biological species are not pests and most pest species do not cause significant harm at all times and in all locations (Pedigo & Higley, 1997). This conceptual breakthrough implied that limiting pesticide use could lead to conserving natural enemies. Hence, the concepts of economic damage, economic injury level and economic thresholds were developed. Economic damage, economic threshold and economic injury level constitute three basic elements of the economic threshold model (Pedigo, 1999).

Economic Damage (ED) is defined by Stern et al. (1959) as �the amount of injury which will justify the cost of artificial control measures�. Based on Stern et al.�s description, Southwood and Norton (1973) presented the following mathematical expression for the economic damage:

C(a) = Y[s(a)]* P[s(a)] - Y(s)*P(s)          (2.1)

Where: Y = yield, P = price per unit of yield, s = level of pest injury, a = control action [s(a) is level of injury as modified by the control action], C = cost of the control action. Equation 2.1 states that cost of the control tactic equals yield times price when the tactic is applied minus yield times prices without the tactic. Consequently, economic damage begins when the benefits of suppression exceeds cost of control, that is, when C(a) ? Y[s(a)] x P[s(a)] - Y(s) x P(s).

Economic Injury Level (EIL) was defined by Stern et al. as the lowest population density that will cause economic damage. The EIL is the most essential of the decision rules. However, it is a theoretical value that if attained by a pest population, will result in economic damage (Pedigo, 1999). Mathematically, the EIL is expressed as (Pedigo et al. 1986):


Where: EIL = expressed as pest numbers or injury equivalents, C = cost of management activity per unit of production (e.g. $/ha), V = market value per production unit (e.g. $/kg), I = injury units (damage) per insect per production unit (e.g. proportion defoliated/(insect/ha), D = damage per unit injury (e.g. (kg reduction/ha)/proportion defoliated, K = the proportional reduction of the insect population).

Economic Thresholds (ET). Several alternative definitions of the economic thresholds have been developed since 1959 when this term was first introduced by Stern et al. (1959). Intuitively, the concept of economic threshold implies that if the pest population and the resulting damage are low enough, it does not pay to take control measures. In practice, the term �economic threshold� has been used: (i) to denote the pest population level at which economic loss (i.e. pest damage) begins to occur and (ii) to indicate the pest population level at which pest control should be initiated given the cost of control (Davidson and Norgaard, 1973).

Stern (1973) defines the economic threshold as the pest density at which control measures should be used to prevent an increasing pest population from reaching economic injury level. The relationship of the ET to the EIL and action times is shown in Figure 2.1. Another definition given by Glass (1975) defines economic thresholds as the density of a pest population below which the cost of applying control measures exceeds the losses caused by the pests (figure 2.2).

Compared to the EIL, the ET is a practical rule rather than a theoretical one. Economic thresholds are developed taking into account three main factors: the physical damage caused by pests at a given level of infestation, the revenue losses resulting from that damage, and the costs of treatment. Mumford & Norton (1984) presented the following measure for the ET:

PDKq = C                                                (2.3)

Where P = price per unit of yield, D = the loss in crop yield per insect per production unit, K = reduction in pest attack, or proportionate reduction in pest attack associated with a particular control action, q = the level of pest attack
C = the cost of control.

The right hand side of (2.3) represent the cost of control and the left hand side represents benefit of control. From (2.3) the economic threshold is derived as:


At this �break-even� point, the benefit of a particular control action (as defined by 2.3) just equals the cost of this control action (C). The expression (2.4) holds for cases when pest problems have linear damage relationship (Fig. 2.3a). In situations where a threshold relationship occurs, Norton and Mumford (1993) redefined expression for economic threshold as:

q* = T + [C / PDK]                                  (2.5)

where T is the maximum level of pest attack below which losses do not occur (Fig. 2.3b).

The economic threshold as defined by Stern et al. has been considered as an operable decision tool. Presently, the use of thresholds is widespread in pest management. Empirically determined economic thresholds, or in other words �action thresholds�, have proven very useful in reducing the number of pesticide sprays that are necessary for controlling various pests (Arneson & Losey, 1997).  In a monograph published recently on IPM in Albania, Isufi (1997, p. 160) lists action thresholds used for controlling pest in several crops including olives in Albania. Suffice to say, the author provides no information on the methodology used to determine those action thresholds.

Several authors have recognized the limitations of the economic threshold model (Poston et al., 1983; Stefanou, 1984; Pedigo, 1999). The ET and EIL concepts have been criticized (i) for being simple and overlooking the influence of other factors within and outside the crop/pest system, (ii) for not incorporating important features such as interseasonal dynamics, interactions with other pests and natural enemies, pesticide resistance, and environmental costs in the typical economic threshold decisions, (iii) for yielding relatively inappropriate results in case of pest management decisions involving multiple pest attacks and multiple treatments. According to Mumford & Norton (1984), it is this last limitation of the economic threshold model that certain agricultural economists have looked for alternative decision models in pest management.

The Economic Optimization Model

Entomologists and economists approach the decision making process in pest management by focusing on very different aspects of the problem. As described above, the question posed by entomologists is at what point (pest population level) a particular control action must be taken to prevent pest population from rising past the point where the damage it would cause exceeds the cost of the control measure. Unlike entomologists, often economists take the pest population as given and seek to answer the question �what level of control is most profitable for that particular pest density� by employing optimality conditions and specifically the law of diminishing returns. The model assumes perfect knowledge and profit maximization in determining optimal solutions.

Hillebrandt (1960) was the first to employ the concept of marginality in pest control. Using a hypothetical example, she describes a dosage response curve in which the crop yield attributable to pesticide application is related to various doses of pesticides and shows how successive pesticide treatment will lead to diminishing returns. Hillebrandt goes on to show there will be an optimal level of pesticide application in terms of profit maximization, beyond which additional treatment gives a reduced increment of returns.

Following the same theoretical grounds, Headley (1972) established a link between economic optimization and earlier work of entomologists on economic thresholds. Headley defines the economic threshold as �the pest population that produces incremental damage equal to the cost of preventing that damage� (p. 105). The concept of economic thresholds as defined by Headley considers three key components: value of production, control costs and pest population. Referring to Figure 2.4, point 1 represents producer revenue in the absence of pest attack. Point 2 shows the tangency points where the rate of change of the value of production is equal to the rate of change of the cost of control function (marginal revenue equals marginal cost). The value of this rate of change is shown at point 3 in the incremental scale.

Given the definition of economic threshold as the break-even point in which the cost of control is equal to the benefits of control, then, point 4 is the economic threshold (pest population associated with point 2). Thus, unlike the definition of economic thresholds provided by entomologists, the economic threshold as defined by Headley (1972) represents the optimal (net return-maximizing) level of pest population. No other pest attack level provides greater net returns above control costs.

Several authors have attempted to deal with dynamic and spatial aspects involved in pest control. Southwood and Norton (1973), Kogan (1976), and Walker (1977) used real data to illustrate the forms that damage and control cost function can have in practice. Hall & Norgaard (1973) made further improvements to Headley�s theoretical model using simple functions for pest population growth, pest damage, pest control, control costs, and time of application of pesticide treatment. Shoemaker (1979) and Conway et al. (1975) employed a more realistic approach that takes into account the population dynamics of pests. They used dynamic programming to derive optimal solutions for the number and timing of control treatments.

The Decision Theory Model

Mumford and Norton (1993) proposed the application of decision theory to pest control recommendations as a means of countering some of the difficulties of using economic thresholds. These and other authors have stressed the need for thresholds to include the farmers� perceptions of crop/pest loss and a better understanding of how farmers make decisions.

Unlike the two previous models that assume perfect knowledge, the decision theory model considers uncertainty in decision-making process of pest management. Risk and uncertainty in pest management arise due to biological, technical, and economic factors as well as farmers� imperfect knowledge about these factors in a given season.

There are three key components to be considered in decision analysis: (i) the events over which the decision maker has control (alternatives), (ii) the probability of the occurrence of chance events, and (iii) the value of various outcomes.

A better description of this model is provided through pay-off matrices. The pay-off matrix lists projected net returns for different pest management practices and severity of pests. Referring to the pay-off matrix in Figure 2.5, each row shows the damage relationship, that is, the amount of damage associated with different levels of pest attack. In the remaining rows, which involve control action (tactics), each cell (outcome) also includes the cost of control action and its effectiveness in reducing pest attack and damage.

The expected outcome for each action is calculated as a probability-weighted average outcome. For example, expected outcome when no action is taken for controlling pests (first row in Figure 2.5) is:

ER0 = P0CN,0 *  PLCL,0 * PMCM,0 * PHCH,0     (2.6)

Assigning probabilities to each state of nature (level of pest severity) is very important to the outcome of the analysis. Related to this, two approaches have been used to calculate probabilities: one based on objective estimates and the other on subjective estimates (Bosch, 1997). Objective probabilities are based on historical information and reflect the past frequency of occurrence of the states of nature. Subjective probabilities are elicited from expert judgment. Also, the outcome cells in the matrix could be subdivided to account for risk associated with crop prices and other factors (Norton and Mullen, 1994). Choosing the best expected outcome depends on the degree of farmers� risk aversion. Those producers, who are unconcerned with variations in outcomes, that is, risk neutral producers, will choose the option with the highest expected monetary value. However, farmers are often risk averse and being so they are willing to trade off some monetary gains for reduced risk of loss.

Carlson (1970) conducted the economic analysis using a decision theory framework to pest control decision making. He elicited subjective probabilities of a range of losses from peach brown rot, the cost of control, and value of the crop by interviewing peach growers. Information on the effectiveness of five possible treatments for the disease was obtained from field trials. Based on this information, Carlson derived net returns and the expected outcome for each pest treatment and presented the results in a pay-off matrix.

Another model that accounts for the degree of uncertainty in pest management decisions is that provided by Feder (1979). Feder expanded the economic optimization model with fixed damage and a control function by introducing random elements for pest level, damage per pest, and effectiveness of the control measure. The farmers� risk attitudes also were considered to investigate the impact of uncertainty on farmers� decisions regarding pesticide use and the way it affects reaction to various changes. The major policy implication of Feder�s model is that even though pesticide use might reduce uncertainty, information regarding old and new technologies could be a substitute for pesticides, depending on relative costs and availability.

The Behavioral Decision Model

This model is based on the belief that often farmers� pest management decisions reflect their individual perception of the pest problems and not the actual situation. Tai (1977) found that pesticide use per hectare varied much more between farmers than between crops in the same farm, even though pest problems were different. Daku et al. (2000) found a similar relationship from a baseline survey conducted with olive growers in Albania. The survey showed that farmers applying pesticides on grapes were more likely to use pesticides on olives than those farmers who had not sprayed grapes at all.

By developing the behavioral decision model, Norton and Mumford (1983) seek to account for these variations in pesticide use among farmers and try to incorporate their behavior into the pest control decision-making. The model is composed of static and dynamic components.

The static model (Figure 2.6) involves the following elements: (i) pest problem and real options for tackling that problem, (ii) farmer�s perceptions on the pest problem, (iii) farmer�s assessment of expected outcome based on the individual perception of the pest problem and options available, (iv) farmer�s evaluation of expected outcome as related to his own personal objectives, (v) action chosen by farmer following the evaluation. Farmers� perceptions of pest problem, which may not necessarily reflect the real situation, constitute the key element to the whole decision making process.

The dynamic decision model (Figure 2.7) accounts for temporal aspects of pest decisions as well as for the influence of other managerial decisions made by farmer within the farm. As Figure 2.7 shows a decision that will be made in a particular season (t+1) can be influenced by the decision in previous season (t), farmer�s experience, and other decision problems.

The major implication coming from this model is that behavioral characteristics of farmers� decision making concerning pest control need to be recognized and evaluated if appropriate pest management strategies are to be developed and implemented.

The Participatory Model

Since the mid 80�s, the emphasis on participatory approaches to IPM has intensified (Roling and van de Fliert, 1994; Nelson, 1994). Norton et al., (1999) mention several factors that have contributed to the growing interest in participatory IPM: (i) information-intensive practices such as pest scouting or monitoring and use of economic thresholds have not reduced pesticide use in all cases and in some cases have increased it, (ii) pest resistance to many pesticides and pest resurgence have prompted farmers in many cases to increase pesticide dosage in order to achieve adequate pest control, resulting in higher pest control costs, lower income, and greater health and environmental risk, and (iii) traditional research and extension approaches have often failed to address the right issues concerning farmer�s concerns in pest management.

In the participatory model, the IPM programs are viewed as an integral part of the participatory research and extension (R&E) system. The �Farmer-Field-School� (FFS) applied extensively initially in Asia and later in Africa and Latin America (Roling and van de Fliert, 1994) and the participatory IPM (PIPM) applied and refined in several developing countries by the IPM Collaborative Research Support Program (IPM CRSP), (Norton et al., 1999) are the most prominent examples of participatory approaches in pest management decision making.

The participatory IPM is grounded in two assumptions (IPM CRSP, 1997):  (i) that research design, implementation, and evaluation must be a collaborative and interdisciplinary process in which both natural and social scientists are involved and (b) that farmers-first concept of Chambers and others (Chambers et al., 1989) must be extended to include multiple actors involved in pest management decision making such as farm managers, growers� associations, marketing agents, policy makers, and so forth. As Figure 2.8 shows (Norton et al., 1999) the PIPM process includes several steps and activities. It begins by identifying site collaborators and other stakeholders, gathering secondary information, and designing a baseline survey and a participatory appraisal.

Three major activities including baseline survey, participatory appraisal, and field monitoring of pests and beneficial organisms are implemented to establish IPM research priorities. The baseline survey serves to identify farmers� pest perceptions, pest management practices and decision making process, basic socioeconomic characteristics, and other information.

The participatory appraisal includes training of natural and social scientists in participatory methods and helps develop a preliminary assessment of research priorities. In the third activity, while the field monitoring of pests and beneficial organisms is underway, farmers and scientists work together to design, test, and evaluate IPM practices. Further, research output is produced, outreach and information exchange are promoted, and economic and environmental impacts of PIPM process are assessed.

Concluding Remarks

Several models that capture farmers� pest management decision process have been presented in this paper. These models along with budget analysis are useful analytical and practical tools that describe why and how farmers make their pest management decisions and help in assessing economic impacts of IPM research at farm level and beyond. However, when used alone, none of the models presented is sufficient to assess the economic impacts of various pest control strategies. Each model and technique has its strengths and weaknesses. A combination of models often provides useful insights to the farmers' incentives for pest control.  In choosing which model to use for a particular situation, several factors must be considered including: (i) the type of questions to be answered and purpose of analysis (ii) the level of analysis (eg. farm vs. aggregate level); (iii) operational aspects such as availability of data, time, financial and human resources, as well as analytical skills; (iv) the completeness and consistency in terms of the conceptual economic framework and the degree of detail and sophistication.


Arneson, P.A. and J.E. Losey. Integrated Pest Mangement � Lecture Notes, Cornell University, 1997.

Bosch, D. Research Methods in Risk Analysis. Lecture Notes. Department of Agricultural and Applied Economics, Virginia Tech, Blacksburg, Virginia, USA, 1997.

Conway, G.R., G.A. Norton, N.J. Small, and A.B.S. King. �A Systems Approach to the Control of the Sugar Cane Froghopper.� In G.E. Dalton (ed.) Study of Agricultural Systems, London: Appl. Sci.. 1975, p. 193-229.

Daku, L., G.W. Norton, D.G. Pfeiffer, G.C. Luther, C.W. Pitts, D.B. Taylor, J. Tedeschini, and R. Uka. �Farmers� Knowledge and Attitudes Towards Pesticide Use and Olive Pest Management Practices in Vlora, Albania: A Baseline Survey.� IPM CRSP Working Paper 001-1, Virginia Polytechnic Institute and State University, 2000.

Davidson, A. and R.B. Norgaard. �Economic Aspects of Pest Control.� Paper presented at the Conference of Plant Protection Economy sponsored by the European and Mediterranean Plant Protection Organization. Brussels, Belgium, 1973.

Feder, G. �Farm Size, Risk Aversion and Adoption of Technology under Uncertainty.� Oxford Econ. Pap. 32 (July 1980): 263-283.

Headley, J.C. �Defining the Economic Threshold.� In Pest Control Strategies for the Future. National Academy of Sciences, Agr. Board, Washington, D. C, 1972.

Hillebrandt, P.M. �The economic theory of the use of pesticides. Pt. 1. The dosage response curve, the rate of application and the area to be treated.� J. Agric. Econ. 13, 1960:464-72.

Isufi, E. Mbrojtja e Integruar e Bimeve. (Durres, Albania: Plant Protection Research  Institute), 1997.

Kogan, M. �Evaluation of Economic Injury Levels for Soybean Insect Pests.� World Soybean Research, Sept. 1976: 515-33.

Mengech, A.N., K.N. Saxena, and N.B. Gopalan. (ed�s). Integrated Pest Management in the Tropics: Current Status and Future Prospects, John Wiley & Sons Ltd, 1995.

Mumford, J.D. and G.A. Norton. �Economics of Decision Making in Pest Management.� Annu. Rev. Entomol., 1984, 29:157-74.

Nelson, K.C. �Participation, Empowerment, and Farmer Evaluation: A Comparative Analysis of IPM Technology Generation in Nicaragua.� Agriculture and Human Values, Spring-Summer 1994: 109-125.

Norton, G.A. and D.J. Mumford. �Decision Analysis Techniques.� Chapter 3 in G.A. Norton and D.J. Mumford (Eds) Decision Tools for Pest Management. Cab International, 1993.

Norton, G.W., E.G. Rajotte, and V. Gapud. �Participatory Research in Integrated Pest Management: Lessons from the IPM CRSP.� Agriculture and Human Values, 16 (1999): 431-439.

Norton G.W. and J. Mullen. Economic Evaluation of IPM Programs: A Literature Review. Report 448-120. Blacksburg, Virginia: Cooperative Extension, 1994.

Pedigo, L.P. Entomology and Pest Management. 2nd edition. Prentice-Hall, Englewood Cliffs, NJ, 1996.

Pedigo, L.P. and L.G. Higley. �Introduction to Pest Management and Thresholds.� Chapter 1 in L. G. Higley and L. P. Pedigo (Eds.) Economic Thresholds for Integrated Pest Management. University of Nebraska Press: Lincoln and London, 1997.

Pedigo, L.P. �Economic Thresholds and Economic Injury Levels.� Posted in National IPM Network�s website, University of Minnesota, 1999. (

Poston, F.L., L.P. Pedigo, and S.M. Welch. �Economic Injury Levels: reality and Practicality.� Bull. Entomol. Soc. Am., 29 (1983): 29-53.

Roling, N. and E. van de Fliert. �Transforming Extension for Sustainable Agriculture: The Case of Integrated Pest Management in Rice in Indonesia.� Agriculture and Human Values, Spring-Summer, 1994, 96-108.

Shoemaker, C.A. �Optimal Management of an Alfalfa Ecosystem.� In G.A. Norton and C.S. Holling (Eds.) Pest Management. Int. Conf., 1979, p. 161-76. Oxford: Pergamon.

Southwood, T.R.E., G.A. Norton. �Economic Aspects of Pest Management Strategies and Decisions.� in P.E. Geier, L.R. Clark, D.J. Anderson, and H.A. Nix (Eds.) Insects: Studies in Population Management. Canberra: Ecol. Soc. Aust., 1973: 168-84.

Stefanou, S.E. �Pest Management Concepts: Past, Present, and Future.� Farm Economics, July/Aug., 1984, 4pp.

Stern, V.M., R.F. Smith, R. van den Bosch, and K.S. Hagen (1959). �The Integrated Control Concept.� Hilgardia 29 No. 2 (1959): 81-101.

Stern, V.M. �Economic Thresholds.� Annual Review of Entomology, 1973, 259-280.

Walker, P.T. (1977). �Crop Losses: Some Relationship between Yields and Infestation.� Meded. Fac. Landbouww et. Rijksuniv. Gent 42 (1977): 919-26.