Policies and Management Practices for Sustainable Oil Palm - Evidence from Indonesia
von Katrin Rudolf
Datum der mündl. Prüfung:2020-05-12
Betreuer:Prof. Dr. Meike Wollni
Gutachter:Prof. Dr. Bernhard Brümmer
Gutachter:Prof. Dr. Matin Qaim
EnglischChanging consumption patterns and a growing world population have resulted in an increased demand for vegetable oil. Due to its price advantages and suitable chemical properties, in particular the demand for palm oil has increased making it globally the most widely used vegetable oil in 2019. Palm oil production is geographically highly concentrated in Indonesia and Malaysia that alone provide over 85% of the global supply. These countries also harbor important parts of moist tropical forests which are one of the world’s most biodiverse ecosystems and global carbon sinks. While the area under oil palm cultivation has been growing continuously in the last decades, the area under rainforest has declined. Part of this decline can be directly attributed to oil palm expansion. Oil palm induced land use changes are associated with environmental degradation. The repercussions are not restricted to the regional or local scale through e.g. a distortion of water and microclimate regulating services, but have global consequences such as biodiversity loss and carbon emissions. Yet, oil palm expansion has contributed to poverty reduction and income growth for many smallholder farmers. The contrast between negative environmental consequences and positive welfare effects highlights the trade-offs associated with oil palm expansion. The present dissertation aims at identifying policies and oil palm management practices that could mitigate these trade-offs. This thesis comprises three essays. The first two essays focus on policies aiming at biodiversity restoration in oil palm dominated areas. The last essay explores management practices to close yield gaps and to increase soil organic carbon (SOC) in smallholdings. All essays area based on primary data collected from oil palm smallholder farmers in Jambi Province, Indonesia. Jambi Province is one of the hotspots of oil palm expansion and has experienced tremendous land use changes. The planting of native trees inside or at the border of the plantation has been suggested as one way to increase biodiversity in oil palm plantations. However, little evidence exists on which policy instruments are effective to promote native i tree planting in oil palm plantations. The first essay of this dissertation addresses this research gap. The agroforestry adoption literature and focus group discussions suggest that missing information and restricted access to seedling markets represent constraints to tree planting. In case these constraints are binding, providing farmers with access to seedlings and information should increase tree planting adoption. To test this hypothesis, a randomized controlled trial was implemented in 2016 and two interventions were designed. The first provided information on tree planting and management through a manual and a video. The second combined information with input provision. In addition to the video and the manual, farmers received six native tree seedlings for free. We estimate intention-to-treat effects on farmers’ tree planting decision in oil palm plantations with a double-hurdle model. Both interventions significantly increase the number of trees planted per hectare in oil palm plantations in comparison to the control group. Under both interventions, a small share of farmers plant many trees in their plantations. However, free seedling provision in addition motivates many farmers to adopt low intensity tree planting. We also analyze tree survival rates. We find that farm-level tree survival is significantly lower for the farmers who received seedlings for free than for the adopting farmers in the control group and for those only receiving information. Non-correspondence between farmers’ preferences and the seedlings received can be identified as one driver. The effective support of ecosystem services operating at larger scales requires to take on a landscape-level perspective. For biodiversity conservation and the support of related ecosystem services, there is evidence that a critical size of conservation area at the landscape level and its spatial connectivity are of importance. The second essay of this dissertation looks into how payments for ecosystem services (PES) schemes for tree planting need to be designed in order to meet these requirements. PES are compensation mechanisms through which providers of ecosystem services, e.g. farmers putting their land under conservation, are compensated for the resulting costs by the respective beneficiaries. The experimental literature has suggested separate design options to reach a critical size of conservation area, ii and to improve its spatial connectivity. However, evidence is missing on how to simultaneously meet both requirements. To close this research gap, we implemented a framed field experiment in 2018. In a hypothetical scenario, farmers needed to decide whether to keep their land under oil palm cultivation or to switch to a mixed system that integrates other trees. While oil palm cultivation is linked to higher income, the mixed system increases biodiversity. To mirror the landscape-level requirements, biodiversity is only improved if a critical number of farmers plant the mixed system. Moreover, effects are higher if mixed systems are grown on bordering land. To explore the effects of potential PES designs, we analyze two incentive schemes. In the first one, the area threshold, farmers who plant the mixed system are compensated if at least three out of six farmers in a group do so. In the second one, an agglomeration payment, adopting farmers receive compensation if at least three farmers in the group plant the mixed system on bordering land. Our results show that both PES designs increase the likelihood that farmers plant the mixed system in comparison to a baseline scenario without compensation payments. However, once communication is allowed for, the area threshold outperforms the agglomeration payment with regard to environmental effectiveness and overall efficiency, while reaching similar levels of budget efficiency. Tree planting can also support climate regulating functions by absorbing and storing carbon dioxide. Soil organic carbon (SOC) restoration represents another approach to improve the carbon sequestration in oil palm plantations. In addition, it might strengthen the provision of yield supporting soil ecosystem services. Increasing SOC can be achieved by adding organic material, e.g. through mulching with empty fruit bunches (EFB). EFB remain after the oil fruits have been stripped off in the palm oil mills. EFB are rich in nutrients and can be used as organic fertilizer. This can help to increase smallholders’ yields that mostly lie below those of industrial plantations. Several studies have indicated that mulching can be beneficial to soil carbon and yields. However, evidence is mostly derived from field-trial experimental data in industrial plantations. These trials might not reflect the real-world conditions of iii smallholder farmers. This motivates the third essay in which we look at the effects of EFB mulching on SOC contents and yields in oil palm smallholdings. Based on yearly data covering the years 2015 to 2017, we estimate the effect of mulching on yields with a random effects panel data model. Since the adoption to apply mulching might be endogenous, we employ an instrumental variable estimation. Based on soil samples collected in 2017, we analyze the effect of mulching on SOC. Finally, we look into the impact channels and analyze whether and to what extent the effect of mulching on yields operates through SOC increases. Our results support a positive effect of mulching on SOC contents and yields. However, SOC contents do not appear to be a significant predictor for yields. The effect of mulching on yields therefore likely operates through other channels such as direct nutrient addition. This dissertation suggests that policies and management practices could help to mitigate environmental-economic trade-offs involved in oil palm cultivation. In case of soil organic carbon restoration, environmental improvements and income generation might be complements. Our results further indicate that biodiversity-friendly farming practices can be promoted with non-monetary interventions. However, to reach a larger share of the population, subsidized inputs or even monetary incentives might be needed.
Keywords: Biodiversity conservation; Soil organic carbon restoration; Randomized controlled trial; Pathway analysis; Payments for ecosystem services; Framed field experiment; Free input provision; Video-based extension; Technology adoption; Impact evaluation