Search

The impact of crop insurance on changes in input use has attracted much attention by economists. While there are a number of studies on this topic, they frame moral hazard in inputs use in a static model. However, when agricultural producers are forward-looking, they would make input allocation decisions realizing that their decisions would affect their future actual production history. This, in turn, affects the probability and size of future indemnity payments. Thus, moral hazard should be framed in a dynamic input use decision model. We first show theoretically that under certain feasible conditions, a static analysis always results in lower optimal input use when compared to a dynamic one with endogenous actual production history. This is because static models fail to recognize the role of actual production history. We then run numerical simulations using nitrogen application rates as a case study. We find that static models indicate significant reduction in nitrogen use compared to the no-insurance scenario, whereas the dynamic models with a role for actual production history indicate almost no reduction in applied nitrogen. The dynamic analysis not only suggests a near absence of moral hazard, but, for low coverage rates, it results in an optimal nitrogen rate higher than that under the no-insurance scenario. These findings illustrate the importance of recognizing the role of actual production history in mitigating moral hazard possibilities in crop insurance and the dynamic nature of moral hazard in crop insurance.

This research analyzes the motivations and dynamics of small farmer participation in supermarket supply chains in developing countries: why some small farmers join these new markets and continue their participation; why others drop out or decline the relationship from the outset. Drawing on insights from the technology-adoption literature on learning and experimentation, and also on findings from a simple two-period Bayesian model of farmer decisions to participate in a new market, we incorporate measures of neighboring farmers’ experience into the decision model. Results suggest that farmers delay entry to observe their neighbors’ outcomes; we find a negative relationship between the number of neighbor participants in a given period and a farmer’s own decision to enter the supply chain. We find evidence that farmers delay entry for strategic reasons, allowing neighbors to bear the costs associated with a first wave of adjustment to the market, including higher product rejection rates and lower initial annual transactions with supermarkets, compared to revenues from marketing in the traditional way. Results are robust to definitions of neighbors using both administrative and geographic designations. Our results raise questions about the optimal sequence and level of farmers' market participation and exit, which remain largely unexplored in the literature.

International audience; Households' monetary valuation of water quality is a prerequisite for efficient water resource management and the valuation of water quality protection policies. Individuals are commonly questioned about their perception of risk in valuation surveys based on stated-preference methods and revealed-preference methods such as averting-behavior models. These subjective and often discrete measures are commonly used to explain individuals' actions to protect themselves against these risks. Perceptions appear as endogenous variables in traditional theoretical averting-decision models but, quite surprisingly, endogeneity of perceived risk is not always controlled for in empirical studies. In this article, we argue that perceptions have to be treated as endogenous to averting decisions in order to produce accurate and reliable measures of households' valuation of water quality improvements. We present various binary averting decision models featuring an endogenous discrete variable (such as risk perception). In particular, we compare the traditional bivariate probit model with the special regressor model, which is less well-known and relies on a different set of assumptions. In the empirical illustration using household data from Australia, Canada, and France, we study how the perceived health impacts of tap water affect a household's decision to drink water from the tap. Individuals' perceptions are found to be endogenous and significant for all models, but the estimated marginal effect is sensitive to the chosen model. Our empirical application also includes some tests of the special regressor estimator's sensitivity to underlying assumptions.

Farmers may value water quality and user safety characteristics of herbicides as they select among products to obtain weed control. Expenditures per application in the U.S. corn and soybean herbicide markets are explained by several safety characteristics in addition to market and weed control characteristics. The explicit inclusion of safety characteristics in the farm decision model indicates that not all safety aspects of pesticide use are external to farmers. Leaching potential and user toxicity are statistically significant, but their elasticities are small relative to broadleaf and grass weed control efficacy.

In this paper, a trade-off function between net returns and soil quality is developed for farmers in southwestern Saskatchewan using a Markov decision model. The results show that farmers who are concerned with soil levels or stewardship employ chem-fallow more often at lower soil moisture levels and also tend to crop more intensively to conserve soil. The major conclusion is that concern for soil quality, as documented by some researchers, does have practical significance in changing agronomic practices; but, when soil is relatively deep, it takes a fairly substantial concern about soil quality before it is possible to distinguish clearly the agricultural practices of farmers who are truly concerned with stewardship from those who are not.

Appropriate management tools are required for the successful introduction of advanced intensive technology in shrimp mariculture. This paper presents a stochastic dynamic decision model for evaluating the potential of the round pond technology practiced at the Oceanic Institute in Hawaii. The model provides the optimal stocking and harvesting schedules for a shrimp pond using a set of intra- and interseasonal decision rules. These rules are expressed as cutoff revenues when both prices and weights are considered random and as cutoff prices and cutoff weights when only prices or weights are considered random. The pond is not harvested if the current realized revenue is less than the cutoff revenue. The model simulates optimal scheduling using a set of 1986 shrimp prices for the case of random prices. Net profits are approximately doubled using the optimal schedules compared to conventional fixed scheduling schemes. The model also evaluates the economics of controlled environment.

What does a decision-making study tell project planners in developing countries about designing a rural development project for small farmers? The results, summarized here, of a study of farmers' decisions regarding the recommendations of the Plan Puebla should begin to answer this question. The Plan Puebla is an agricultural development project that was started by Centro Internacional de Majoramiento de Maiz y Trigo (CIMMYT) in 1967 and is now run by the secretariat of agriculture in Puebla, Mexico. Its aim is to increase yields of maize on rainfed farms. Initially, the project's breeding program attempted to find improved maize varieties or hybrids that performed appreciably better than the local variety (criollo). When the program found none, the project focused on deriving recommendations about fertilizer use and timing and plant population for the local variety (CIMMYT). The aim of this study was to view the Plan Puebla through the eyes of the proposed adopters of the new technology-the farmers. The study focused on the decision-making processes of a small sample of farmers in one village concerning the recommendations of the plan in order to identify the factors limiting adoption of the recommendations. The recommendations for the village in 1973-74 were to get credit for fertilizer, to increase plant population, to increase the number of fertilizer applications, and to use a recommended level of fertilizer per hectare. Since farmers seemed to decide to adopt each of the first three recommendations independently of the other two, separate models of the first three adoption decisions were developed. In order to identify the factors limiting adoption of the fourth recommendation, a simulation model of the quantity of fertilizer per hectare demanded by individual farmers was developed. Since farmers first adopted fertilizer in 1962, the mode focuses not on the decision to adopt fertilizer, which occurred long enough ago to cause problems in modeling and data collection, but on the decision to increase or decrease fertilizer use since then. Before particular decision models and results are presented, a discussion of the form of the decision model used is in order. Recently, social scientists have argued that people, in choosing one alternative over other available alternatives, do not make complex calculations of the overall utility of each alternative. Rather, people tend to use procedures that simplify their decision-making calculations (Simon, Cyert and March, Tversky, Quinn). Hierarchical models or trees, with decision criteria at the nodes or branching points of the tree, can represent such procedures. The tree form of the decision model is the product of a theory of the choice process that takes into account the simplifying procedures people use in "real-life" decision making.

Previous agricultural pest management information studies have focused on the consequences of providing pest information to farmers without access to other sources of pest information. In view of the proliferation of commercial scouting services, this paper uses a preposterior decision model of pest management to examine the outcomes likely to follow from introduction of public pest information into an area with an existing commercial pest information market. Empirical estimates of the impact of a soybean insect forecast system in Illinois demonstrate that although farmers can benefit from public pest information, the implications for pesticide use depend on the reliability of public forecasts.

Extension is under increasing pressure to conduct evaluations that provide substantive and reliable evidence of program results. However, current evaluation methods which can supply these results are too costly for general applications. This study proposes an alternative evaluation method which measures program participants' belief changes and uses normative decision models to estimate the effects of behavioral responses to those belief changes. A pilot application demonstrates that the proposed evaluation procedure is a viable alternative to current methods, with potential cost advantages that could permit more widespread application.

Sri Lanka rice farmers are faced with a choice between traditional and high-yielding rice varieties. Three risky-choice models are compared to explain producer behavior. Forty farmers with typically small rice areas were studied. Choice criteria compared were multiple (two)-attribute utility maximization, single (profit)-attribute utility maximization, and expected profit maximization. Expected utility maximization performed better than expected profit maximization; single-attribute utility maximization outperformed the more complex multiattribute utility model. The importance of careful elicitation of subjective probability distributions is emphasized. Rice farmers appear to allocate their land between available technologies in an economically rational way.

Historic estimates of livestock production functions have used repeated observations on the animals to estimate the overall gain surface. The estimates involve autocorrelation accordingly. To circumvent this problem, gain isoquants are estimated directly by means of an instrumental variable approach. Data are from a swine experiment designed for the purpose with 648 hogs. Protein supplement and corn serve as the substitute inputs. Protein supplement is estimated as the instrumental variable. Isoquant equations then are predicted for three weight intervals. Finally, a decision model is applied for economic choice. Comparisons are made for various price sets of recent years.

The flexible moment-based approach to production analysis is used to measure the stochastic structure of large-scale dairy production in Tulare County, California. This approach is flexible because it does not impose restrictions on the relationships between decision variables and the moments of the probability distribution of output. The parameter estimates support the hypothesis that second and third moments are functions of inputs and indicate that capital-intensive dairies are riskier. The implied behavior of a risk-averse dairy manager changed markedly when the effects of inputs on the skewness of the output distribution were included in a decision model.

A number of states are currently developing comprehensive programs for making computerized decision aids available to farmers (Harsh; Infanger, Robbins, Debertin; LaDue; Schoonaert; Walker). While Cernea and Tepping have outlined procedures for a general evaluation of agricultural extension projects, only a few efforts have been conducted with a corresponding research effort aimed at quantifying the impacts of computerized decision models on farmer behavior. (Erickson; Erickson, Sandborg, Byers; Knoblauch). This paper summarizes results from a pilot research project designed to provide measurements of changes in farmer behavior as the result of being exposed to a computerized planning model in extension. This study is an attempt to determine the specific management decisions that the farmers intended to make as a result of having been exposed to the planning model. A follow-up effort one year later identified managerial decisions that actually had been implemented as a result of exposure to the planning model a year earlier. Hence, both farmers' intentions and realizations were measured in this

Despite being a pillar of modem economic thinking, expected utility theory (EUT) has many limitations for explaining behavior in the presence of uncertainty. These limitations are well known and oft repeated (e.g., Arrow, Runge and Myers), but economists have just begun to pursue other theoretical frameworks for understanding how both market and nonmarket institutions cope with real-world uncertainty. The potential rewards of this pursuit are ample since EUT's contribution has been at the individual decision-making level, whereas the most serious theoretical deficiencies are in accounting for the linkages from individual decisions to institutional formation and behavior (Coleman, p. 184). The merit of the pursuit notwithstanding, uncertainty is a difficult analytical concept because it has been confused with discussions of risk and subjective probability distributions. To further the confusion, many writers have resorted to pretentious adjectives such as "profound" and "irresolvable" (Goodin) or "structural" (Walters) to distinguish true uncertainty. In the spirit of Knight, the clearest definition of uncertainty is lack of knowledge of states or actions and, at the extreme, sheer ignorance.1 Uncertainty created by imperfect knowledge and humans' limited reasoning skills is the most dominant consideration in understanding any organization or living system (Hodgson, Kay, Walters). One key problem for any theory of institutional behavior under uncertainty is to explain the origin and stability of institutions rather than simply to treat them as exogenous. Neoclassical decision models based on EUT have limited value for this task because the market institution is the centerpiece of analysis. Other social institutions are viewed as constraints, and information problems are defined as transactions costs that only impact the allocation from exchange (e.g., Dahlman, Schmid, Williamson). As a result, the influence of uncertainty on the origin and functions of these institutions is not fully explored.2 The recent work of Heiner (1983, 1985, 1986) on reliability theory (RT) offers an alternative theoretical framework to explain institutions as responses to problems of uncertainty. Given that institutions are a "regularity in social behavior" (Schotter, p. 11), RT is promising because it "directly harnesses the determinants of uncertainty to characterize regularity in behavior" (Heiner 1983, p. 571). In the following section, we outline the basic principles of RT and contrast it with EUT. We then explore the implications of RT for the The authors are associate professors in the Food and Resource Economics Department, University of Florida. Florida Agricultural Experiment Station Journal Series No. 9215.

The pesticide application practices of California peach growers in controlling peach brown‐rot are used to demonstrate how Bayesian decision theory procedures can be used to arrive at optimal crop disease control practices. Subjective probabilities of disease loss intensity are measured and used in the decision model. Information from an analyst (this researcher) is combined with farmers' subjective probabilities of disease loss by means of Bayes' theorem. Optimal pesticide use actions are computed for three different objective functions—maximum subjective expected returns, mean‐standard deviation of returns, and maximum expected returns with a minimum income side condition.

PROFIT maximization over time is the goal of economic production regardless of whether the production process is of fixed or variable length. The classical criteria for profit maximization are consistent with maximization over time when the production process is of fixed length. Variable length production processes, as encountered in livestock feeding, require somewhat different criteria from those given by the classical theory of the firm. The length of a process can affect both production costs and revenues (e.g., length of feeding of livestock). When there is a time lag between the end of one variable length process and the beginning of another, profit maximization is achieved by extending the process so long as the extension results in additions to revenue greater than or equal to additions to cost.

Purchase of grain dryers by farmers can be explained using a threshold theory of decision in which a reaction occurs only after the strength of the stimuli increases beyond the farmers' reaction threshold. Estimates of the causal relationships in this dichotomous decision model were obtained by using the probit technique with data from a farmer panel. The results provide an explanation of differences in the number of dryers purchased by farmers in various classes and regions. It also provides a procedure for predicting dryer ownership and purchases of dryers as an aid to forward planning by distributors of drying equipment.

This article presents the conceptualization of a farm firm growth model which includes both investment and production decisions. The results obtained in the application of the model to a hypothetical Central Indiana farm under alternative resource availability and optimizing criteria assumptions are reviewed. These applications demonstrate the interaction between investment and production decisions in the growth process and indicate that under realistic conditions optimal firm growth is not scale growth. They also indicate that the growth process and the terminal size are different for different objective functions. Finally, the results illustrate the applicability of such a model for investigating farm firm adjustments in a growth context.

A PLICATION OF Bayesian decision theory to problems facing agriculture has come during a decade of continually increasing interest in the weather [12]. The National Weather Service spends about 70 million dollars annually observing, analyzing, and forecasting weather phenomena. In addition to the usual general forecasts, specialized forecasts of particular value in local areas are used in agricculture as well as other industries [13, p. 86]. In the future forecasts of specific weather phenomena will undoubtedly become increasingly available, either as a public service from the Weather Service or for sale by private consulting firms. The public forecasting service would want to avoid the cost of issuing forecasts that cannot profitably be used by the clientele in the forecast area. Or farmers purchasing private forecasts would want to avoid paying a price that exceeds their value. Thus, as the capability of issuing forecasts increases, it seems reasonable to assume that agricultural economists will be called upon to estimate their value. The Bayesian decision model would seem to be suited to the task of estimating the value of weather forecasts. The model itself has been extensively developed in the literature [5, ch. 5 and 6; 9; 11, ch. 5; 14; 15]. Its use in price forecasting has been demonstrated by Eidman, Dean, and Carter [7] and Bullock and Logan [1]. Carlson [4] illustrated its use for crop disease forecasting. Byerlee [2] and Byerlee and Anderson [3] have used it to evaluate a long-range weather forecast. Estimates of the production function, the prior probability distribution for weather inputs, and the likelihood function are needed to compute the value of a forecast. The production function would include the random weather variables to be forecast as well as a set of inputs under the manager's control. It would be used to compute the profits from alternative actions and suggest the form of the forecast. Unfortunately, when using Bayesian analysis the information needed to estimate the value of the forecast is the same information needed to

The theory of personal, or subjective, probability has been applied in many areas of decision analysis but has had limited use in agricultural economics (Carlson, Bessler, Francisco and Anderson). This limited use of subjective probability theory is curious since considerable potential exists for applying it to supplement the limited data available for supporting agricultural decision models and assessing the consequences of extension and other programs designed to modify perceptions of an identifiable clientele. Recently, Grisley and Kellogg (GK) have used the theory of subjective probability in eliciting price, yield, and net return distribution estimates for several crops from farmers in Thailand. Although their analysis is illustrative of the potential for the theory, several decisions were made in the application which have unfortunate consequences for their results. These choices in the application indicate a basic confusion regarding the use of scoring rules in eliciting subjective probability distributions (SPDs) and appropriate methods for assessing the "goodness" of the elicited SPDs. The purposes of this note are to identify explicitly the problems in the GK application and to suggest a few general guidelines for future applications.

O NE of the critical problems faced by those who contemplate investments is the effect of taxes and finance charges on investment decisions. While it is known that these costs reduce the expected annual return available for retiring the investment cost, there is at present no readily applicable method of taking this known reduction into consideration in the decision process. This paper develops a method by which taxes, investment credit, and financing charges can readily be included in the decision model. A more meaningful economic appraisal of investments results. At the outset the requirement is made [1] that