Applying Utility Theory in Agricultural Decision-Making

Every farming decision – whether to plant a drought-resistant variety, purchase crop insurance, or try a new high-value crop – involves choosing between uncertain outcomes. Some choices offer higher potential rewards but carry greater risk of loss; others are safer but limit upside. How should a farmer rationally navigate these trade-offs? Utility theory offers a structured, mathematically grounded answer. Rooted in economics and decision science, it gives farmers and agribusiness managers a framework for making consistent, preference-aligned choices even when the future is unknown.

What is utility theory? 🔗

Utility theory is a framework used to model decision-making under uncertainty. Rather than simply maximizing expected monetary returns, it accounts for the fact that individuals have preferences – and that the same outcome can feel very different depending on a person’s financial situation, goals, and attitude toward risk. In the context of agriculture, utility is understood as the personal satisfaction or well-being a farmer derives from a particular outcome. Two farmers facing identical choices may make entirely different decisions because their utility – and their tolerance for risk – differs.

The formal treatment of utility in decision-making traces back to the mathematician Daniel Bernoulli, who in 1738 argued that it is the utility of gains, not the monetary value alone, that drives rational choice. This laid the groundwork for what later became the expected utility hypothesis – the principle that a rational decision-maker will choose the option that maximizes the expected value of their utility function, not just expected income.

The expected utility hypothesis 🔗

The expected utility hypothesis holds that when a decision-maker faces a set of risky outcomes, they will choose the option that maximizes the weighted sum of utilities across those outcomes, where the weights are the probabilities of each outcome occurring. Formally, if a farmer faces outcomes O₁, O₂, …, Oₙ with corresponding probabilities P₁, P₂, …, Pₙ and utility values U₁, U₂, …, Uₙ, the expected utility (EU) is:

EU = P₁·U(O₁) + P₂·U(O₂) + … + Pₙ·U(Oₙ)

The farmer is expected to choose the option with the highest EU. Critically, this is not the same as choosing the option with the highest expected monetary value. A farmer might reject a crop variety with a high average expected income if the variance (and thus the risk of catastrophic loss) is too great – because the utility loss from a bad outcome outweighs the utility gain from a good one.

According to Britannica, the implication of the expected utility hypothesis is that consumers and firms seek to maximize the expectation of utility rather than monetary values alone, and since utility functions are subjective, different people can approach the same risky event with very different valuations.

The von Neumann-Morgenstern utility index 🔗

The theoretical backbone of utility theory in modern economics is the work of mathematician John von Neumann and economist Oskar Morgenstern. In their landmark 1944 work, Theory of Games and Economic Behavior, they moved beyond subjective notions of satisfaction and proved that any rational decision-maker whose preferences satisfy a set of logical axioms must have a utility function – now called the von Neumann-Morgenstern (VNM) utility function or VNM utility index.

The VNM theorem demonstrates that rational choice under uncertainty involves maximizing the expected value of a cardinal utility function. It proves that if a decision-maker’s preferences are consistent with four behavioral axioms – completeness, transitivity, continuity, and independence – then there exists a utility function that numerically represents those preferences, and the person will always act to maximize its expected value.

The four axioms behind the VNM index 🔗

The construction of the VNM utility index rests on four axioms that define rational preference behavior:

• Completeness: A farmer can rank any two outcomes – they either prefer one over the other, or are indifferent between them. There is no “I can’t decide” scenario.
• Transitivity: Preferences are consistent. If a farmer prefers outcome A over B, and B over C, they must prefer A over C. Violating this leads to what economists call a “money pump” – a situation where the farmer would keep paying to swap between options irrationally.
• Continuity: There is always some probability at which a farmer is indifferent between a certain outcome and a gamble involving the best and worst outcomes. This ensures the utility function is smooth and mathematically workable.
• Independence: Preference between two options is not affected by mixing both with a third common outcome at the same probability.

When these axioms hold, the VNM theorem guarantees the existence of a utility function unique up to positive affine transformations – meaning the scale and origin of the function can shift, but the preference rankings it generates remain the same.

How the VNM utility index is constructed 🔗

To construct a farmer’s VNM utility index in practice, the analyst begins by identifying the best possible outcome (B) and the worst possible outcome (W) the farmer could face. These are assigned utility values – typically U(B) = 1 and U(W) = 0.

For any intermediate outcome Y (such as a moderate harvest income), a probability P is found at which the farmer is indifferent between receiving Y with certainty versus entering a lottery that yields B with probability P and W with probability (1-P). The farmer’s utility for Y is then simply U(Y) = P. According to the VNM framework, the utility of any intermediate outcome can be found as a linear transformation of the two extreme outcomes.

By repeating this indifference-probability elicitation for different income levels, a complete utility function can be mapped – one that numerically encodes the farmer’s risk attitude. This function is the VNM utility index.

Risk attitudes and their meaning in agriculture 🔗

The shape of a farmer’s utility curve directly reveals their attitude toward risk. The VNM utility function can be used to explain three distinct profiles: risk-averse, risk-neutral, and risk-seeking behavior.

Risk-averse farmers 🔗

A risk-averse farmer has a concave utility function – meaning each additional unit of income adds less utility than the previous one (diminishing marginal utility). For such a farmer, the disutility of losing ₹10,000 is greater than the utility gained from earning ₹10,000. They place higher utility on the expected value of a gamble than on taking the gamble itself.

This is by far the dominant pattern in agriculture. A study from Odisha, India found that nearly 65% of surveyed farmers exhibited risk-averse behavior. Research on Polish farmers similarly found strong risk aversion, with crop insurance emerging as the most important risk-coping strategy. Risk-averse farmers are more likely to choose traditional low-variance crops, buy crop insurance, and avoid technologies with uncertain outcomes.

Risk-neutral farmers 🔗

A risk-neutral farmer has a linear utility function. They evaluate outcomes purely by expected monetary value, and are indifferent between a sure payment and a gamble with the same expected payout. These farmers make decisions based on expected income alone, making them well-served by accurate market data and price forecasts.

Risk-seeking farmers 🔗

A risk-seeking (or risk-loving) farmer has a convex utility function and actively prefers higher-variance gambles – they place greater utility on taking the gamble than on its expected value. This profile is less common in agriculture but does appear. Research from Zambia found that farmers exhibiting risk-seeking behavior tended to have lower wealth, suggesting that some risk-taking may be driven by economic desperation rather than preference for volatility.

Applying utility theory to real farm decisions 🔗

The VNM utility index and expected utility hypothesis are not just theoretical constructs – they translate directly into practical decision-making tools for farmers and agribusiness managers.

Crop choice under weather uncertainty 🔗

Consider a farmer in a drought-prone region deciding between two crops. Crop A is a traditional variety with moderate and stable yield: ₹60,000 income whether or not there is moderate rainfall. Crop B is a high-value commercial crop: ₹1,20,000 income in a good year, but only ₹15,000 in drought. Historical data suggests a 40% probability of drought.

The expected monetary value of Crop B (0.6 × ₹1,20,000 + 0.4 × ₹15,000 = ₹78,000) is higher than Crop A. But a risk-averse farmer – one whose utility function is concave – would calculate the expected utility of each crop, not just expected income. If the utility loss from a very poor harvest is large (and for a small farmer depending on that harvest for food and loan repayment, it likely is), the expected utility of Crop A may exceed that of Crop B. The expected utility framework explains why the farmer rationally chooses the lower-income but more stable option.

Crop insurance decisions 🔗

Utility theory is central to understanding why farmers do or do not purchase crop insurance. Research on Ghanaian cocoa farmers shows that risk-averse farmers are significantly more willing to participate in crop insurance programs, as they are willing to trade off production risks in exchange for assured indemnification. A risk-averse farmer’s utility function means the expected utility with insurance (lower but more stable income) exceeds the expected utility without insurance (higher expected income but with the risk of catastrophic loss).

Studies on Italian grape growers found that crop insurance effectively causes risk-averse farmers to behave as if they were risk-neutral, enabling them to invest optimally in inputs rather than under-investing out of precaution. This is a powerful practical consequence of the utility framework: insurance realigns the farmer’s decision from risk-adjusted utility maximization to straightforward profit maximization.

Technology adoption 🔗

Utility theory also explains patterns in the adoption of new agricultural technologies. Research from Zambia found that risk aversion was significantly associated with reduced adoption of conservation agriculture practices such as minimum tillage and crop rotation. When the utility loss from a failed experiment is high – particularly for resource-poor farmers – the expected utility calculation favors sticking with proven methods, even if the expected monetary payoff of the new technology is higher.

Limitations of utility theory in agriculture 🔗

While utility theory provides a powerful normative framework, it has real-world limitations. Behavioral research shows that real decision-makers consistently violate the axioms of expected utility theory – particularly the independence axiom. Farmers’ actual choices are often influenced by framing effects, loss aversion, and ambiguity (where probabilities themselves are unknown, not just outcomes). This has led to the development of alternative frameworks such as Prospect Theory and Cumulative Prospect Theory, which better describe actual behavior.

Empirical studies in agricultural economics have found that expected utility theory lacks descriptive accuracy in certain contexts, as risk preferences can shift over time. That said, the VNM framework remains the dominant tool for applied agricultural risk analysis, decision modeling, and farm policy design – and even its critics acknowledge it as the indispensable baseline from which all other models depart.

What do you think? If most farmers are risk-averse by nature, does this mean that agricultural policies should automatically prioritize stability over efficiency – or is there a case for encouraging calculated risk-taking in farming? And how might the utility framework need to evolve to better capture the decisions of smallholder farmers in low-income countries, where the stakes of a bad harvest go far beyond financial loss?

References

1. https://www.fao.org/4/w7365e/w7365e0e.htm
2. https://link.springer.com/chapter/10.1007/978-1-349-20568-4_11
3. https://www.britannica.com/topic/von-Neumann-Morgenstern-utility-function
4. https://en.wikipedia.org/wiki/Von_Neumann%E2%80%93Morgenstern_utility_theorem
5. https://www.encyclopedia.com/social-sciences/applied-and-social-sciences-magazines/utility-von-neumann-morgenstern
6. https://pmc.ncbi.nlm.nih.gov/articles/PMC7066235/
7. https://ideas.repec.org/a/ags/stagec/196907.html
8. https://pmc.ncbi.nlm.nih.gov/articles/PMC10901030/
9. https://onlinelibrary.wiley.com/doi/10.1111/1467-8489.12434
10. https://www.mdpi.com/2227-7099/10/1/20
11. https://ageconsearch.umn.edu/record/271520/
12. https://www.academia.edu/23799179/A_theoretical_and_empirical_examination_of_Expected_Utility_Theory_and_Prospect_Theory_in_agricultural_economics_literature
13. https://nimss.org/projects/view/mrp/outline/18604