Evaluating a Payments for Ecosystem Services program in Uganda

Seema Jayachandran, Associate Professor of Economics Northwestern University, Affiliated Professor and Health Co-Chair, the Abdul Latif Jameel Poverty Action Lab (J-PAL),  discusses the effectiveness of payments for ecosystems (PES) for reducing deforestation in developing countries.

Finding effective ways to curb deforestation is a key question for climate policy: Deforestation and forest degradation account for about 20 percent of anthropogenic carbon emissions, most of which occur in developing countries. Slowing deforestation is also a potentially very cost-effective way to reduce emissions compared to more expensive interventions in transport and energy. The UN has already invested millions in REDD and REDD+ to help curb deforestation. These programs may be incorporated into global carbon markets under the next international climate treaty, resulting in billions of dollars in payments from rich to poor countries for forest conservation.

Despite these investments, there is little rigorous evidence about the most effective ways to reduce deforestation in developing countries. Take the case of payments for ecosystem services (PES). In a PES program, participants receive payments if they follow a set of conditions that are protective of the environment, such as not cutting down trees on their land. PES is a popular policy approach in many developing countries. It is perceived to be a relatively simple intervention compared to other conservation programs and one that could generate new sources of financing for conservation.

Yet, surprisingly few programs have been carefully evaluated. Past evaluations of PES programs have relied on anecdotal evidence or measuring forest cover before and after the program, making it impossible to learn to what extent a change in forest cover is the result of the program or the result of other factors. Do payments for forest conservation change landowners' behavior, socio-economic well-being, and increase forest cover? To help answer these questions, Joost de Laat, Eric Lambin, Charlotte Stanton, and I are conducting a randomized evaluation of a PES program in Uganda, which began in 2011.

Over the past decade, the use of randomized controlled trials (RCTs) to evaluate the effectiveness of social programs has grown dramatically, particularly in the fields of education, health, and microfinance. RCTs are widely regarded as the most credible and reliable way to measure the impact of a program. By randomly assigning a group of people to either receive an intervention (the treatment group) or not (the control group), we can conclude that the difference in outcomes between the groups will be the result of the intervention itself and not external or unobserved factors.

Why randomize in the case of PES? Like other conditional cash transfers, PES schemes involve paying people for something they may have done even without the program. Forest owners who would have conserved their forests anyway are likely to sign up for a PES contract at a much higher rate than forest owners who planned to cut down trees. Thus, comparing those who take up the program and those who do not would vastly overstate the benefits of PES: Much of the estimated "benefit" of the program would be due to selection, not actual policy impact. Similarly, a simple before and after comparison could dramatically overstate the effect of PES if, during the time period it was implemented, the deforestation was declining for other reasons too; conversely, researchers would underestimate the benefits of PES if the rate of deforestation was increasing. The random assignment of treatment and control villages in our study vastly improves on these methods by providing a valid counterfactual for assessing the policy impact that leads to a clear measure of the additional forest cover generated by the PES program.

Forest loss in Uganda is estimated to be about 1 to 2 percent per year, with an even higher rate on private land. Our evaluation is based in western Uganda in Hoima and Kibaale districts, which have some of the highest deforestation rates in the country. Forest owners often cut trees to clear land for growing cash crops such as tobacco and rice or to sell the trees as timber or for charcoal production. In our study, we randomly assigned 136 villages to either receive a PES program or serve in the comparison group. In the 65 villages randomly assigned to receive the program, our local partner, the Chimpanzee Sanctuary and Wildlife Conservation Trust (CSWCT) offered an incentive contract to each individual landowner. Under this contract, landowners receive annual payments if they refrain from cutting trees on their land and re-forest a portion of their land. CSWCT employees conduct random spot checks to monitor compliance by checking for newly cleared patches of forest or fresh tree stumps.

One challenge in conducting RCTs is knowing whether the program under evaluation generates benefits to society as a whole or just redistributes existing benefits. In the case of PES, if landowners stop cutting down their own trees, they could simply switch to cutting down trees on other land. Thus, PES could replace deforestation in one location with deforestation in another. To get at this question, our study will also measure the spillovers onto land not covered by the program on nearby government forest reserves.

The PES program and our evaluation are still ongoing. Our early analysis shows that there is low take-up of the PES contract. In my next post, I'll share some of our preliminary analysis on the reasons behind low take-up and discuss why landowners' difficulty obtaining loans when they need emergency cash may limit the effectiveness of PES programs.