UC Santa Barbara

This dissertation consists of three essays. The first aims to compare the cost-effectiveness between command-control and market instruments in addressing non-point source pollution. By definition, non-point source pollution (NPSP) is extremely difficult to observe individual level discharge and thus, very hard to implement market incentive policies. I exploit a policy setting where agricultural runoff is in fact, a point source pollution but is regulated as if it were NPSP which allows the study of abatement behavior in what is typically a NPSP setting. I study a program called the Florida Everglades Forever Act intended to reduce phosphorus runoffs from entering the sensitive Everglades ecosystem. The program consists of both a command-control component as well as a market incentive component which I am able to disentangle using a new dataset I developed on annual farm level discharge and subsidies for pollution reduction. I use the two-step Arellano-Bond estimator to estimate a marginal abatement cost (MAC) curve for the average farm. With the estimated MAC curve, I simulate the costs under the market mechanism and compare that with both data-estimated and engineer-estimated costs under command-control. I find that to achieve the same benchmark pollution outcome, the market mechanism would reduce compliance cost by 20%.

The second chapter examines the theoretical efficacy of an ambient mechanism in ameliorating the NPSP problem. Specifically I examine theoretically how an ambient mechanism to ameliorate the NPSP problem can produce free-riding incentives. Specifically, I show the conditions in which uncertainty about firm types may lead to incorrectly setting the uniform ambient tax rate which then creates the potential for free-riding. I also compare the Nash and Sub-game Perfect Nash equilibria and analyze the potential welfare gains of adding more water quality monitoring points. I find that expanding the network of water monitors in such a setting does not always reduce free riding potential compared to the single-monitor case though it never rises above this level. The reason is that splitting the group by adding more monitors could simply be redistributing the free-riding potential to the multiple groups rather than actually decreasing the free-riding potential of all groups together.

Chapter three is joint work with Chris Costello which discusses the role of indemnity in Payments for Ecosystem Services programs (PES). PES programs are voluntary programs where private or public beneficiaries of ecosystem services (a public good) agree to pay private producers of ecosystem service (ES) inputs. However, when there is private risk to the private provisioning of ES inputs, then there may be gains to offering loss protection (indemnity). This paper characterizes conditions in which it is optimal for a budget constrained regulator to (i) offer indemnity in conjunction with a linear pricing contract and (ii) to pursue the dual objective of poverty alleviation and maximizing social benefits from ES inputs. We find that it is optimal for the regulator to share in the risk of producing ES inputs (or outputs), i.e., offer full indemnity if agents are risk averse. Furthermore, the value from optimally choosing the indemnity, compared to the no-indemnity case, is higher whenever agents are more risk averse and can lead to as much as a 40% increase in ES supply for the same budget. We also provide a guide to practitioners and empirical researchers on how to evaluate the appeal of indemnity in any particular setting for which PES exists and provision of which is risky. Lastly, we identify a estimatable threshold for the business-as-usual ES supply curve slope above which it is optimal to pursue the dual objective.