Essays on incentives and pro-environmental behaviour

This thesis consists of four self-contained essays at the nexus of applied microeconomics, behavioural economics, and environmental economics. In the essays of the thesis, I use field experiments and econometric tools to examine the impact of monetary and non monetary incentives for behavioural change during resource scarcity. I use methods of eliciting intrinsic motivations and then empirically test theories on the interaction of intrinsic motivation and extrinsic incentives. Specifically I analyse whether and which incentives undermine, support, or are independent of existing preferences, and whether incentives change behaviour. The first two essays analyse two distinct types of conservation policy in Brazil: i) direct payments from the state of São Paulo to small landholders living in vulnerable ecosystems conditional on the landholders conserving their land; and, ii) federal policy to regulate, monitor and enforce land use in the Brazilian Amazon through conservation zoning and creation of a public list of municipalities with high rates of deforestation (“priority municipalities”) to increase visibility and thus accountability. The first essay¹ uses a field experiment in Brazil to test if monetary incentives to conserve land on private property in vulnerable ecosystems - “Payments for Environmental Services” (PES) - crowd out demand for a conservation program. Landholders are less likely to accept the higher monetary offers to conserve compared to the lowest offers. Given that the rational choice model does not explain the role of incentives in shaping demand for PES, we then look at the interaction of the randomised incentive offers and individuals’ initial intrinsic motivations. We construct methods to elicit social preferences in order to analyse this interaction. We find that, while high monetary incentives crowd in demand of progovernment landholders, they crowd out demand of pro-environment (henceforth “proenvironment”) and prosocial landholders. The second essay² combines satellite data on deforestation with data on the location and timing of the conservation zones in Brazil to estimate the effect of conservation zoning on deforestation in the period 2004-2010. We provide spatial regression discontinuity estimates and difference-in-difference estimates to show that the policy does not explain the large reduction in deforestation rates during this period. We provide evidence that zones reduce deforestation in municipalities put on a federal government “shame” list for high deforestation rates. The last two essays³ test behavioural interventions to decrease residential water consumption across the City of Cape Town in South Africa as complements to tariff increases and water restrictions during a severe water crisis. Using inserts in monthly municipal bills, we test multiple behavioural messages in a randomised control trial on the full population of free standing domestic households (400 000+). The treatments are classified into five groups: information provision and increased salience on the tariff structure, financial savings, appeals to the public good, social comparison, and social recognition. By using a number of different framings, the third essay focuses on identifying which incentives best motivate individuals of different income levels to reduce their consumption. We find that lower income households respond only to financial incentives, whereas the higher income households respond only to social incentives and appeals to their intrinsic motivation. In the final essay, we further explore the drivers behind the effect of social recognition on pro-environmental behaviour (henceforth “proenvironment behaviour”). According to Bénabou and Tirole (2006), the visibility of doing-good may create doubt to others as to the true motive of the individual and result in a crowding out of prosocial behaviour. We use three treatments within the larger randomised control trial to disentangle intrinsic motivation, extrinsic incentives, and image motivation. We exogenously vary the visibility of the social recognition treatments to test whether i) social recognition incentives crowd out intrinsic motivation and, ii) whether social recognition increases the noise of the prosocial signal and ultimately crowds out cooperation. We find, on average, using image motivation as an extrinsic incentive crowds in cooperation. Social recognition with an explicit opt-out has, on average, no effect on consumption. Thus, in our setting, the signal of social recognition for prosocial behaviour is strong enough to elicit cooperation. In application to public policy, our findings suggest public recognition can be used as an adjunct to more traditional demand side management tools, such as water restrictions and tariff increases to achieve additional conservation in the higher income households. To our knowledge, this empirical analysis has not been executed elsewhere and contributes both to the academic literature as well as policy recommendations for alternatives to traditional demand side management tools during times of resource scarcity. ¹Co-authors: Florence Kondylis, Development Research Group, World Bank; Astrid Zwager, Development Research Group, World Bank. ²Co-authors: Liana O. Anderson, National Center for Monitoring and EarlyWarning of Natural Disasters (CEMADEN); Torfinn Harding, Department of Economics, NHH Norwegian School of Economics, and University of Stavanger; Karlygash Kuralbayeva, Grantham Research Institute, LSE; Andre Lima, Department of Geographical Sciences, University of Maryland. ³Co-authors: Kerri Brick, Environmental Policy Research Unit (EPRU), University of Cape Town; Martine Visser, School of Economics, University of Cape Town.