Achieving landscape patterns for biodiversity conservation through payments for ecosystem services – Evidence from a field experiment in Indonesia

Abstract

Payments for ecosystem services (PES) can encourage land owners to manage their land in a biodiversity-friendly way. To increase the effectiveness of PES for biodiversity conservation, incentives could be set to reach a minimum size of conservation area and a suitable spatial connectivity between conserved areas. However, little knowledge exists about which incentives most likely generate such landscape patterns since threshold effects and spatial connectivity have been mostly considered separately. Therefore, we present results from a framed field experiment with Indonesian oil palm farmers and compare the effectiveness between two conditional group payment schemes: In the first one, the size threshold payment, payments are made if at least three farmers in the group conserve. In the second one, an agglomeration payment, payments are made if at least three farmers with bordering land engage in conservation. Our results suggest that both PES designs are similarly effective in the absence of communication. Under both, communication increases conservation outcomes in the case of previous successful coordination and is ineffective in the case of previous coordination failure. Yet, for individuals who are reluctant to conserve, communication only increases conservation outcomes under the size threshold payment. We further discuss potential welfare implications of our results.

Introduction

Biodiversity loss is at its highest level ever seen and species extinction rates are accelerating, as indicated by the 2019 Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services report (Diaz et al., 2019). To set incentives for private land owners to manage their land in a biodiversity-friendly way, which can help to stop species loss, payments for ecosystem services (PES) have been suggested. While PES have become a widely implemented policy instrument in both developed and developing countries, the associated benefits for the environment appear to be mixed (Börner et al., 2017). This could be, among others, because the design of PES schemes often concerns the natural resource management at the plot level only, and neglects the surrounding landscape. This is problematic because landscape characteristics can be important for the protection of various species and the generation of ecosystem services (Goldman et al., 2007). In this study, we experimentally investigate the performance of two alternative PES designs that take landscape characteristics into account.

Landscape characteristics relate to the composition and configuration of the landscape, i.e., the proportion of different habitat types and their spatial arrangement (Goldman et al., 2007). For instance, a minimum size of conservation area can be critical for species richness and abundance, such as for the protection of various forest-dependent bird species (Betts et al., 2007). Regarding landscape configuration, large distances between, and isolation of forest patches negatively correlate with the existence and abundance of several species (Martensen et al., 2012). Besides evidence from forest areas, the relevance of critical levels of biodiversity-supporting land uses and of connectivity has also been supported for semi-natural, conserved or restored areas in agricultural landscapes (Hass et al., 2018; Azhar et al., 2011; Tscharntke et al., 2011). Notwithstanding these findings, empirical evidence on how to design incentives to achieve critically-sized, spatially connected landscape patterns remains scarce.

In the literature, several PES designs have been suggested to promote landscape patterns for biodiversity conservation, including size threshold payments, agglomeration bonus schemes, and agglomeration payments. Size threshold payments aim at generating critically sized conservation areas, and therefore payments are made to providers of ecosystem services only if the required size is generated at a predefined scale (e.g. Midler et al., 2015; Narloch et al., 2012). Agglomeration bonus schemes and agglomeration payments are designed to increase the spatial connectivity between conserved areas. The agglomeration bonus is paid in addition to a uniform base payment, e.g. per hectare conserved, if farmers put bordering land under conservation (Parkhurst et al., 2002). The agglomeration payment, in contrast, is only paid if farmers put bordering land under conservation, i.e., no uniform base payments are offered under this scheme (Wätzold and Drechsler, 2014; Drechsler et al., 2010). Previous studies have found that size threshold payments effectively increase the provision of threshold environmental goods that are only generated if a critical conservation area size is present (e.g. Midler et al., 2015; Narloch et al., 2012). Furthermore, Parkhurst et al. (2002) have shown that agglomeration bonus schemes can effectively increase spatial coordination between land owners. Simulation studies on agglomeration payments have highlighted possible increases in the provision of ecosystem services in comparison to designs without an explicit spatial target or agglomeration bonus schemes (Wätzold and Drechsler, 2014). While these studies support that PES can be designed to achieve landscape patterns for conservation, they assume that the generation and size of environmental benefits depend on either reaching a critically sized conservation area or on its connectivity, but not on both.¹ For biodiversity conservation, however, both landscape characteristics can be relevant.

In this study, we analyze the effects of two different PES designs on the generated environmental benefits using a framed field experiment. We add to the existing literature by considering that the generated environmental benefits in our experimental set-up depend on both a critical area size and spatial connectivity, as can be relevant for biodiversity conservation. The two PES designs vary with respect to the implied conditionality. The first one is a size threshold payment that compensates providers of ecosystem services for conservation if at least three individuals conserve (e.g. Midler et al., 2015). The second one is an agglomeration payment that compensates providers of ecosystem services for conservation if at least three individuals with bordering land conserve. Thus, the latter scheme is spatially explicit and provides additional incentives for spatial coordination. To explore the relevance of the size of the compensation payment on conservation efforts, we combine the two PES designs with three different payment levels.

Empirical evidence on whether real-world decision makers coordinate under spatially explicit PES schemes, such as agglomeration bonus or agglomeration payment schemes, is scant.² While studies on size threshold payments have used framed field experiments with e.g. farmers in Peru (Midler et al., 2015; Narloch et al., 2012), evidence on the effectiveness of agglomeration bonus schemes is mostly derived from laboratory experiments with general students (Banerjee, 2018; Banerjee et al., 2017). Evidence on the effects of agglomeration payments is almost entirely derived from simulation studies (Wätzold and Drechsler, 2014; Drechsler et al., 2010). These simulations assume that farmers will put bordering land under conservation, if expected revenues are positive. Coordination failure, which has often been observed in laboratory experiments (e.g. Banerjee, 2018), is hence ignored. A notable exception is the study by Ferré et al. (2018), which experimentally investigates the effectiveness of agglomeration payments with agricultural students, but requires coordination among two individuals only.

Our study is implemented in Indonesia, a country that has experienced a rapid expansion of oil palm monocultures, associated with unprecedented biodiversity losses (Vijay et al., 2016). In the framed field experiment, Indonesian farmers decide whether to keep their land under oil palm monoculture, or whether to switch to conservation, i.e., a biodiversity-friendly mixed system that integrates trees besides oil palms. We thus contribute to the scarce literature integrating spatial aspects in framed field experiments with real-world decision-makers. Especially for the case of agglomeration payments, which have mostly been investigated with simulation studies, we generate experimental evidence that can complement the results from simulations and increase external validity.

In our experimental set-up, the two PES designs and three different payment levels are first introduced without communication. While in real-world villages, individuals usually communicate at least to some degree, communication can be seriously inhibited, e.g., if villagers belong to different ethnicities or religions, or if part of the land is owned by absentee landholders who live in urban areas (Jelsma et al., 2017). We then analyze how the introduction of communication affects the performance of the two PES designs and explore whether its effects on individual conservation choices depend on previous coordination experience. Our study is structured as follows. The next section provides background on the case study. Section 3 describes the design of the framed field experiment and derives hypotheses regarding farmers' behavior in the experiment. Section 4 presents the results and section 5 discusses potential implications for the generated environmental benefits and efficiency, which are implied by the model parameters. Section 6 concludes.