Commodity buyers increasingly pledge to adopt zero-deforestation sourcing and to improve their smallholder suppliers’ livelihoods. They typically offer positive incentives—price premiums, subsidized inputs, training—either individually, to smallholders who do not clear forest on their own plots, or collectively to an entire local community, if no one clears forest in an area. We propose another alternative: reward a community if no one derives economic benefit from clearing forest in the area. We embed this area no-use condition, together with the common individual and area no-deforestation conditions, in a game-theoretic model of heterogeneous farmers who can form intra-family coalitions, clear forest, and, at a cost, block others from using cleared plots. We formalize a novel “Pessimistic Recursive Core” solution concept that enables us to analyze our game with partial cooperation and multiple non-cooperative equilibria, and to check when a conditional incentive prevents deforestation (possibly even compensating farmers for their opportunity costs). We find that the best design hinges on local conditions. With full cooperation, area no-deforestation is cheapest and most robust. With limited cooperation, area no-use dominates at low blocking costs, but risks inequity, whereas otherwise individual contracts are best, but these require perfect incentives (each smallholder individually preferring the incentive to clearing forest). Increased cooperation can either raise or lower the area no-use condition’s effectiveness. Calibrated to 58 oil-palm villages we surveyed in East Kalimantan, Indonesia, the framework indicates that village-specific premiums tied to the area no-use rule prevent deforestation in most cases and are resilient to external encroachment.

Increasing the welfare of smallholder farmers in developing countries plays a crucial role in the global effort to reduce worldwide poverty and hunger. On the one hand, smallholders represent a large proportion of the world’s poor and, on the other, they produce the majority of the food consumed in developing countries. This realization has led governments and organizations around the world to implement policies aimed at increasing farmers’ yields. Although most of these policies have resulted in welfare increases, the environmental effects have been varied. While in many settings intensification policies have been linked to a decrease in deforestation, in many other settings the reverse is true. In this chapter we propose a novel explanation of these seemingly contradictory results. We achieve this through studying a detailed operational model of a farmer’s dynamic decisions of land-clearing and production. We show the importance of considering the interaction between random production costs and liquidity constraints faced by smallholder farmers. These two elements are key to our main result: a reduction in the cost of intensification can lead to lower deforestation rates when the variation in production costs is high enough compared to the cost of intensification. Alternatively, the same reduction in the cost of intensification may lead to higher deforestation rates if the variation in production costs is low enough compared to the cost of intensification. This result helps explain the discrepancies seen in practices and may allow policy makers to better target interventions in order to achieve win-win situations: improvement of smallholder welfare and protection of the natural forest.

Effective preparation for future pandemics requires a clear understanding of how best to deploy non-pharmaceutical interventions, especially population confinements. Experience during COVID-19 shows that many jurisdictions tailor confinements by population group or by activity, yet such targeting is operationally demanding and politically sensitive, making rigorous cost–benefit quantification indispensable. We develop a modeling framework in which confinements can be targeted along two dimensions—age group and activity—to minimize a composite loss that combines mortality and foregone economic output. A stylized, analytically tractable version of the model yields closed-form optimal confinement rules and conditions that reveal when targeting generates welfare gains, and clarifies how these gains depend on key epidemiological and economic parameters. We find that targeting yields gains in certain L-shaped parameter regions, and gains behave non-monotonically in parameters. To translate these insights into practice and quantify their impact, we introduce a structured optimization procedure that couples model-predictive-control techniques with trust-region methods to derive high-quality solutions. A full-scale implementation for COVID-19 in \^{I}le-de-France demonstrates that targeting by age or by activity delivers Pareto improvements relative to non-targeted, uniform policies; and targeting along both dimensions delivers Pareto improvements relative to targeting along just one. We extend the model and the algorithm to deal with ambiguity in problem parameters through a distributionally-robust approach; we find that gains from targeting persist under parameter ambiguity and surprisingly, more ambiguity can increase these gains. Applying a structured optimization approach to derive optimized targeted confinements can therefore be highly beneficial even in the early stages of a pandemic, when estimates of epidemiological parameters are unreliable.

Many operational settings share the following three features: (i) a centralized planning system allocates tasks to workers or service providers, (ii) the providers generate value by completing the tasks, and (iii) the completion of tasks influences the providers’ welfare. In such cases, the planning system’s allocations often entail trade-offs between the service providers’ welfare and the total value that is generated (or that accrues to the system itself), and concern arises that allocations that are good under one metric may perform poorly under the other. We propose a broad framework for quantifying the magnitude of value losses when allocations are restricted to satisfy certain desirable guarantees to the service providers. We consider a general class of guarantees that includes many considerations of practical interest arising (e.g., in the design of sustainable two-sided markets) in workforce welfare and compensation, or in sourcing and payments in supply chains, among other application domains. We derive tight bounds on the relative value loss and show that this loss is limited for any restriction included in our general class. Our analysis shows that when many providers are present, the largest losses are driven by fairness considerations, whereas when few providers are present, they are driven by the heterogeneity in the providers’ effectiveness to generate value; when providers are perfectly homogenous, the losses never exceed 50%. We study additional loss drivers and find that less variability in the value of jobs and a more balanced supply-demand ratio may lead to larger losses. Lastly, we demonstrate numerically using both real-world and synthetic data that the loss can be small in several cases of practical interest.

Given a simple graph G, a set 𝐶⊆𝑉(𝐺) is a neighborhood cover set if every edge and vertex of G belongs to some G[v] with 𝑣∈𝐶, where G[v] denotes the subgraph of G induced by the closed neighborhood of the vertex v. Two elements of 𝐸(𝐺)∪𝑉(𝐺) are neighborhood-independent if there is no vertex 𝑣∈𝑉(𝐺) such that both elements are in G[v]. A set 𝑆⊆𝑉(𝐺)∪𝐸(𝐺) is neighborhood-independent if every pair of elements of S is neighborhood-independent. Let 𝜌_n(𝐺) be the size of a minimum neighborhood cover set and 𝛼_n(𝐺) of a maximum neighborhood-independent set. Lehel and Tuza defined neighborhood-perfect graphs G as those where the equality 𝜌_n(𝐺′)=𝛼_n(𝐺′) holds for every induced subgraph 𝐺′ of G. In this work we prove forbidden induced subgraph characterizations of the class of neighborhood-perfect graphs, restricted to two superclasses of cographs: 𝑃₄-tidy graphs and tree-cographs. We give as well linear-time algorithms for solving the recognition problem of neighborhood-perfect graphs and the problem of finding a minimum neighborhood cover set and a maximum neighborhood-independent set in these same classes. Finally we prove that although for complements of trees finding these optimal sets can be achieved in linear-time, for complements of bipartite graphs it is NP-hard.

Environmental science documents that agricultural production is a dominant driver of deforestation in developing countries. Complex land tenure systems coupled with costly enforcement of conservation laws enable farmers to convert tropical forests into productive land in search of a better income. In this paper, we model the incentives and specific decision-making processes of farmers who dynamically choose their land expansion in response to specific economic and value-chain circumstances. We analyze several interventions that balance forest protection and farmers’ welfare.