top of page

A Game of Spillovers

Conservation Strategies in Pemba and their Unintended Consequences


Pemba Island, located just off the coast of Tanzania, is only 67 km long and 22 km across, yet continued deforestation there has drawn considerable attention and funding from international institutions over the last few decades. Global conservation interventions, such as the Reduced Emissions from Deforestation and Land Degradation or REDD+ program funded largely by the Royal Norwegian Embassy, aim to reduce global carbon emissions through forest protection and by supporting the economic development of local communities. A REDD+ readiness initiative was implemented in Pemba in 2015.

The program promised an environmentally and economically sustainable future for the island, where 90 per cent of households rely exclusively on forest products to meet their daily needs. REDD+, however, failed to produce any tangible reduction in forest cover loss on Pemba and did not deliver the promised carbon credit payments for efforts to slow deforestation. On combining insights from Pemba with current research, one can see how conservation interventions have become part of the lived experience for forest-dependent communities targeted for REDD+.

Conservation deluge

The great, yet predictable, irony of climate change is that the people who are most harmed live in the world’s most impoverished communities — the populations least responsible for greenhouse gas emissions. While discrete, catastrophic events such as the deadly European heatwave in 2019 and the recent Texas snowpocalypse in 2021 garner much of the public attention in the West, the mundane and continuously creeping impacts of climate change are devastating rural livelihoods in the developing world. Consequences like sea-level rise and coastal erosion are disastrous for people who depend on natural resources to meet their daily needs. Even as the global economy trends towards low-carbon futures, the climatic shifts we see today will worsen for the remainder of this century.

Our best strategy for protecting rural livelihoods unequivocally includes conserving a broad array of natural areas. Natural areas buffer human communities against flooding, erosion, and other consequences of our changing environment, while actively sucking up atmospheric carbon. With the recognition that corporations from the Global North have irreparably damaged the world’s ecosystems over the last half-century, we now invest $24 billion on conservation annually to “offset” our bad behaviour.

The number and size of geographic areas designated for conservation are skyrocketing each year. The general strategy for land conservation is to buy and set aside areas that still have most of their native vegetation, face high development pressure, and are cheap to purchase. The deluge of conservation dollars into these areas is juxtaposed with the fact that they largely overlap with the very populations most affected by climate change; populations who depend on forest products for food, fuel, building materials, and cultural needs. Thus, the day-to-day needs of local people and externally-led conservation projects inevitably collide.

Institutions like REDD+ that operate on the private carbon credit market fund conservation and allow countries and corporations to realize their net-zero emissions goals. Interventions are carried out in specific locations for specific lengths of time. It is relatively easy to calculate the amount of carbon captured by limiting local access to a designated area for a number of years and how much it will cost to incentivize communities to forgo those resources.

Wicked spillovers

The issue with this model is that conservation is a game of spillovers. Interventions regularly have consequences outside of the initial place and duration of the project. These can be beneficial. Marine protected areas, for example, give safe harbour to fish populations where they can multiply, leave, and increase yields for fishing boats far outside their boundaries. On land, parks have upfront costs to establish but quickly bolster tourism economies in neighboring towns. It may come as no surprise, though, that these spillovers are not always beneficial for local people or the environment.

One particularly wicked conservation spillover is leakage. Leakage occurs when interventions limit harvest or land development in a specific location, but harvest pressure is only displaced onto adjacent landscapes, resulting in no net reduction in resource loss. Leakage is a perennial problem in land conservation. It has limited the efficacy of oil palm certification programs in Borneo by incentivizing palm expansion into other agricultural areas. Efforts to limit agricultural sprawl in the United States resulted in a 25 per cent increase in development pressure after the conservation contract. Forest conservation aimed at developing an eco-tourism economy in Costa Rica pushed an increase in forest development outside park boundaries. Leakage into neighboring Cambodia considerably attenuates the countrywide successes seen in forest regrowth in Vietnam. And finally, efforts to curb the high profile land conversion in the Peruvian and Brazilian Amazon both resulted in increased deforestation in neighboring forest areas.

In forest-dependent communities, social and ecological processes are inherently interwoven; changes in the environment affect peoples’ behaviour, which in turn affects the environment and future behaviours in ways that can be difficult to predict. Interventions that work well in one place may have drastically different outcomes in another. On Pemba Island, for example, the REDD+ program promised communities funding for development projects in exchange for slowing deforestation rates. As nearly every household in Pemba depends on daily forest access to meet their cooking needs, reducing access to local resources forced community members to shift their harvest effort onto neighboring areas. Leakage in this case made it difficult to quantify the amount of carbon captured as a result of the program. The managing organisation terminated the intervention before delivering the promised community payments and did not succeed in reducing forest cover loss. However, the programme did succeed in causing notable tensions between communities and seeding distrust in conservation projects.

For people who depend on daily access to natural resources, the long-term and large-scale outcomes of conservation interventions matter. Yet, economically driven, conservation-for-carbon interventions can be indifferent to downstream social outcomes. Systematic reviews show that leakage and other unintended consequences are getting worse and happening more often as conservation accelerates worldwide and climatic instability is making them more challenging to predict.

Mitigating conservation externalities

Conservation is necessary for protecting rural livelihoods against the increasingly inhospitable consequences of climate change. These interventions will likely always alter the lifestyles of nearby communities in various ways. The challenge is then, to better document, predict, and mitigate these externalities. Moreover, unaccounted for externalities which make conservation untenable for local people affect how receptive residents will be to future interventions, limiting our overall capacity for conservation.

Long-term and large-scale consequences of conservation are not beyond us. We can monitor these more distal spatial-temporal effects, not just on biodiversity or carbon storage, but also on public opinion. Further, it is communities themselves who are best positioned to observe these nuanced social-ecological outcomes, not external institutions. Community involvement in interventions is, in fact, one of the most influential drivers of conservation success. Research has demonstrated that local communities may better predict the outcomes of conservation interventions than can foreign practitioners. Long- term, community-led monitoring is the first step to understanding and addressing the spillovers that plague current conservation efforts.


Matt is a PhD student in the Human-Environment Systems group at Boise State University. He works in collaboration with the Max Planck Institute for Evolutionary Anthropology to study how conservation projects are affecting landscapes and livelihoods in Pemba Island, Tanzania. Matt is interested in using math and computers to figure out how to do conservation better.

Art by Ben Beechener.


bottom of page