📈 economics

Tackling water theft: an empirical framework for compliance-inducing penalties under uncertainty

This paper presents an empirical framework demonstrating that effective deterrence of water theft requires penalties significantly higher than current sanctions, driven primarily by enforcement credibility and agricultural profitability rather than fixed environmental damage assessments.

Original authors: Giammauro Soriano, Noureddine Bouzidi, Manuel Bea Martinez, Carlos Dionisio Perez Blanco

Published 2026-10-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Giammauro Soriano, Noureddine Bouzidi, Manuel Bea Martinez, Carlos Dionisio Perez Blanco

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Water theft is a silent crisis draining the world's rivers and aquifers, accounting for an estimated thirty to fifty percent of the global water supply. It happens when farmers or industries pump water without a legal right or beyond the limits of their permits, often to grow lucrative crops. For decades, authorities have tried to stop this by fining offenders, but these penalties have largely failed. The core problem is a mismatch in incentives: the money a thief makes from stealing water is often far greater than the cost of the fine they might pay. If the risk of getting caught is low and the fine is small, stealing water becomes a rational business decision rather than a crime. To solve this, regulators need to understand exactly how much money a thief stands to lose if they are caught, and how likely they are to be caught, to set a penalty that makes theft unprofitable.

A team of researchers has now built a new way to calculate these necessary penalties, moving beyond simple guesses to a rigorous, data-driven approach. They focused their study on the Doñana region in southwestern Spain, a critical sanctuary for migratory birds and a UNESCO World Heritage site that is currently under severe threat. The area is famous for its strawberry and berry production, which generates hundreds of millions of euros in export revenue. However, this agricultural boom has come at a steep ecological cost. Over-pumping and illegal wells have reduced the water flowing into the marshes to less than twenty percent of what it was in the 1950s. Despite repeated warnings from international bodies and a landmark ruling by the European Court of Justice condemning Spain's failure to protect the area, illegal irrigation continues to expand, covering thousands of hectares.

The researchers set out to determine why current fines are failing and what it would actually take to stop the theft. They combined satellite imagery, which can spot illegal irrigation even under plastic tunnels or in low-lying crops, with detailed field data to map exactly where the water is being stolen. They found that in the study area, about 20.75 million cubic meters of water are being used illegally every year. This illicit water supports crops like olives, berries, and strawberries, generating an estimated gross profit of 15.91 million euros for the farmers using it. The team then used complex economic models to simulate the behavior of farmers in different towns across the region. They tested thousands of scenarios where crop prices, yields, and water availability changed, asking a simple question: at what point does the cost of getting caught become higher than the profit from stealing?

The results were stark. The current fines in Doñana range from 0.29 to 1.35 euros per cubic meter of water stolen. The researchers' simulations showed that to effectively deter theft, the penalty would need to be much higher. The median penalty required to make farmers comply was found to be 3.01 euros per cubic meter. In many scenarios, the necessary penalty soared even higher, reaching averages of over 27 euros per cubic meter. This means that the fines currently being imposed are so low that they cover only a tiny fraction of the potential profit a thief could make. Even the highest current fine of 1.35 euros would fail to deter two-thirds of the illegal water use in the region. The study suggests that as long as the profit from theft exceeds the expected cost of the penalty, farmers will continue to steal water, regardless of how many satellites are watching.

The analysis also revealed that the size of the fine is not the only, or even the most important, factor. The researchers discovered that the likelihood of a thief actually being prosecuted and convicted is the dominant driver of compliance. In the current system, while detection rates might be high thanks to modern technology, the chances of a case moving through the courts to a guilty verdict are extremely low. The study found that if the probability of prosecution and conviction were higher, the required fine to stop theft would drop significantly. Conversely, if the legal system remains slow and ineffective, no amount of increasing the fine on paper will work, because the expected cost of breaking the law remains low. This creates a situation where the penalty is effectively a "cost of doing business" rather than a deterrent.

The researchers argue that the current legal framework, which bases fines on the environmental damage caused rather than the profit gained from the theft, is fundamentally flawed. In Doñana, proving the exact amount of environmental damage is difficult and often leads to reduced fines in court. The study suggests a shift toward a system where penalties are designed to strip away the illicit profits entirely, a concept already being explored in other parts of the world. However, the authors caution that simply changing the penalty calculation is not a magic bullet. Their simulations show that without fixing the institutional bottlenecks that prevent prosecutions and convictions, even the most advanced penalty systems will struggle to succeed. The economic incentive to steal water is too strong to be overcome by awareness campaigns or voluntary agreements alone; the expected cost of non-compliance must be raised to match the expected gain.

Ultimately, the study provides a clear roadmap for policymakers. It shows that effective water governance requires a dynamic approach where penalties are adjusted based on market conditions and enforcement credibility. As climate change makes water scarcer and crop prices fluctuate, the value of stolen water will rise, and static fines will become even less effective. The path forward involves not just setting higher fines, but ensuring that the legal system is capable of enforcing them. By aligning the economic reality of the penalty with the reality of the profit, and by strengthening the institutions that deliver justice, it may be possible to turn the tide on water theft and protect vital ecosystems like Doñana from further degradation.

Technical Summary: Tackling Water Theft – An Empirical Framework for Compliance-Inducing Penalties Under Uncertainty

Problem Statement
Water theft, defined as the use of water without rights or beyond established limits, accounts for 30–50% of the global water supply and is increasingly prevalent in agricultural regions. Current enforcement policies, particularly penalties based on environmental damage rather than illicit profits, have largely failed to deter illegal abstraction. This ineffectiveness stems from a failure to account for the adaptive responses of economic agents to changing socioeconomic, environmental, and institutional conditions. Traditional damage-based models often allow offenders to treat penalties as a "cost of business," especially when the profits from illegal irrigation significantly exceed the expected cost of non-compliance (penalty multiplied by detection, prosecution, and conviction probabilities). While Loch et al. proposed a conceptual shift toward profit-based penalties, their framework lacked empirical application due to data limitations and the computational complexity of testing multiple equilibria under uncertainty.

Methodology
This study develops and applies a replicable empirical framework to estimate compliance-inducing penalties under uncertainty, integrating multi-model ensemble economic modeling, Global Sensitivity Analysis (GSA), and remote sensing data.

Case Study: The framework is applied to the Doñana region in southwestern Spain, a UNESCO World Heritage Site suffering severe ecological degradation due to illegal irrigation. The study focuses on 13 municipalities covering approximately 61% of the region's illegally irrigated land (5,647 hectares).
Data Integration:
- Remote Sensing & Field Data: A 1:5,000 scale land use map was generated using Sentinel-2 satellite imagery, 0.25m aerial orthophotography, and cadastral data. Irrigation detection utilized spectral indices, visual analysis, and field validation (achieving 99.7% reliability for plastic-covered crops and 91% for low-coverage woody crops). Infrastructure markers, such as solar panels linked to pumping stations, were used to identify irrigation in low-vegetation crops.
- Economic Data: Crop prices, yields, and costs were sourced from ECREA.
- Enforcement Data: Probabilities of detection, prosecution, and conviction were derived from administrative records and legal precedents, noting a historically low conviction rate (2.2–4.4%).
Modeling Approach:
- Positive Mathematical Programming (PMP): An ensemble of three PMP models (Standard, Average, and Paris calibrations) was calibrated at the municipal level to simulate irrigator behavior. These models maximize gross margins subject to land and water constraints.
- Penalty Calculation: The compliance-inducing penalty is calculated as the foregone profit (gross margin reduction) resulting from the elimination of illegal water use, scaled by the joint probability of enforcement (detection × prosecution × conviction).
- Global Sensitivity Analysis (GSA) & SHAP: A Latin hypercube design sampled seven inputs (price, cost, yield multipliers, legal water availability, and enforcement probabilities) across 9,216 scenarios. A random-forest emulator (trained with $R^2$ = 0.99) was used to generate SHAP (SHapley Additive exPlanations) values, quantifying the contribution of each input to the variability in required penalties.

Key Results

Penalty Gap: The median compliance-inducing penalty required to deter illegal water use is 3.01 EUR/m³, with a mean of 27.23 EUR/m³ due to a right-skewed distribution. This substantially exceeds the currently observed penalties in Doñana, which range from 0.29 to 1.35 EUR/m³. The study estimates that current upper-bound penalties would fail to deter approximately 66.6% of illegal water abstractions.
Drivers of Deterrence: SHAP analysis reveals that prosecution and conviction probabilities are the dominant drivers of compliance-inducing penalties (contributing ~34.9% and ~33.6% respectively), followed by crop prices (13.8%) and yields (11.1%). Detection probability contributed less (0.8%) because it was sampled over a narrow, high range (0.80–1.00) reflecting assumed systematic monitoring.
Non-Linearity: The relationship between enforcement and penalties is inverse and non-linear. As prosecution and conviction probabilities increase, the required monetary penalty drops rapidly before approaching an asymptote. Conversely, low institutional enforcement probabilities necessitate hyperbolic increases in penalties to maintain deterrence.
Economic Impact: The 20.75 million m³ of illegal water annually generates an estimated gross margin of EUR 15.91 million. Strawberry production alone accounts for 50% of this illicit profit.

Significance and Claims
The paper claims to provide the first empirical implementation of a compliance-inducing penalty framework under uncertainty, addressing the gaps identified by Loch et al. regarding the testing of multiple equilibria.

Structural Transition: The results support a structural transition from damage-based penalties to profit-based regulatory frameworks. The authors argue that fixed, damage-based penalties are unlikely to remain effective under changing socioeconomic and environmental conditions because they do not align the cost of non-compliance with the illicit gains.
Institutional Bottlenecks: A critical finding is that even with advanced detection technologies (e.g., satellite monitoring), penalties are neutralized if prosecution and conviction probabilities remain low. The study mathematically demonstrates that institutional bottlenecks dilute the expected cost of non-compliance, rendering high nominal fines ineffective.
Complementary Governance: The authors posit that economic instruments (penalties) are a necessary but not sufficient condition for deterrence. They argue that economic frameworks should be complemented by behavioral interventions and institutional strengthening (e.g., supply-chain traceability, linking subsidies to compliance) to address the root incentives for illegal behavior.
Dynamic Thresholds: The study concludes that the compliance-inducing penalty is not a static legal quantity but a dynamic economic threshold that evolves with market conditions, water scarcity, and institutional performance. Therefore, regulatory systems must adopt adaptive strategies rather than static penalty schedules.

The paper does not claim that its framework solves water theft entirely but rather provides a quantitative method to identify the specific deterrence thresholds required under various plausible scenarios, highlighting the critical interplay between agricultural profitability and enforcement credibility.

Technical Summary: Tackling Water Theft – An Empirical Framework for Compliance-Inducing Penalties Under Uncertainty

More like this