​Visual Inference and Graphical Representation in Regression Discontinuity Designs (with Carl Lieberman, Jordan Matsudeira, Zhuan Pei and Yi Shen) [link] [working paper version]

Quarterly Journal of Economics 138 (3) 2023: 1977–2019

Despite the widespread use of graphs in empirical research, little is known about readers’ ability to process the statistical information they are meant to convey (“visual inference”). We study visual inference within the context of regression discontinuity (RD) designs by measuring how accurately readers identify discontinuities in graphs produced from data-generating processes calibrated on 11 published papers from leading economics journals. First, we assess the effects of different graphical representation methods on visual inference using randomized experiments. We find that bin widths and fit lines have the largest impacts on whether participants correctly perceive the presence or absence of a discontinuity. Our experimental results allow us to make evidence-based recommendations to practitioners, and we suggest using small bins with no fit lines as a starting point to construct RD graphs. Second, we compare visual inference on graphs constructed using our preferred method with widely used econometric inference procedures. We find that visual inference achieves similar or lower type I error (false positive) rates and complements econometric inference.

Promoting Spatial Coordination in Flood Buyouts in the United States: Four Strategies and Four Challenges from the Economics of Land Preservation Literature (with Polina K. Dineva, Kent D. Messer, Leah H. Palm-Forster, Laura A. Paul and A.R. Siders) [open access link]

Natural Hazards Review 24 (1) 2023: 05022013

Managed retreat in the form of voluntary flood-buyout programs provides homeowners with an alternative to repairing and rebuilding residences that have sustained severe flood damage. Buyout programs are most economically efficient when groups of neighboring properties are acquired because they can then create unfragmented flood control areas and reduce the cost of providing local services. However, buyout programs in the United States often fail to acquire such efficient, unfragmented spaces, for various reasons, including long administrative timelines, the way in which buyout offers are made, desires for community cohesion, and attachments to place. Buyout programs have relied primarily on posted price mechanisms involving offers that are accepted or rejected by homeowners with little or no negotiation. In this paper, we describe four alternative strategies that have been used successfully in land-preservation agricultural-environmental contexts to increase acceptance rates and decrease fragmentation: agglomeration bonuses, reverse auctions, target constraints, and hybrid approaches.We discuss challenges that may arise during their implementation in the buyout context—transaction costs, equity and distributional impacts, unintended consequences, and social pressure—and recommend further research into the efficiency and equity of applying these strategies to residential buyout programs with the explicit goal of promoting spatial coordination.

​​Choice Uncertainty and the Endowment Effect (with Steven G. Otto) [pdf]

Journal of Risk and Uncertainty 65 (1) 2022: 83-104 

We experimentally test for the role of choice uncertainty in generating “endowment effects” - the robust empirical finding that endowing participants with an item raises their valuation relative to participants being asked to purchase it instead. While there is some compelling evidence concerning trade uncertainty in the literature, there is substantially less evidence regarding the importance of choice uncertainty. This paper provides novel support for the significance of choice uncertainty in the context of both trading and stated valuations. We find that reducing choice uncertainty eliminates under-trading in the exchange setting and decreases (but does not eliminate) the difference in average valuations reported by buyers and sellers, mainly by decreasing the number of extreme valuations by sellers. Interestingly, our treatment does not lead to a significant increase in the number of mutually acceptable trades implied by stated valuations. Comparing the results from our two experiments therefore suggests that value uncertainty continues to play a role in generating valuation asymmetries even after relevant product uncertainty has been resolved. 

Who Will Pay for Increasing Biofuel Mandates? Incidence of the Renewable Fuel Standard Given a Binding Blend Wall (with Harry de Gorter and David R. Just)

American Journal of Agricultural Economics 101 (2) 2018: 492–506

We show that the cost of increasing biofuel mandates given a binding ethanol blend wall falls disproportionately on diesel fuel consumers. The extent of the burden on diesel fuel consumers is explained neither by their relatively more inelastic demand nor by blenders seeking to capitalize on the biodiesel tax credit. Relaxing the blend wall constraint by increasing the potential demand for high-ethanol blends is the only effective lever to insulate diesel fuel drivers from the one-sided welfare impacts of rising mandate levels. The independent effects of the nested mandate structure and the joint compliance base under the Renewable Fuel Standard (RFS) generate the link between motor gasoline and diesel fuel markets. Our results highlight the importance of evaluating the incidence of the RFSin a holistic framework taking both ethanol and biodiesel into account.

Demystifying RINs: A Partial Equilibrium Model of US Biofuel Markets (with David R. Just)

Energy Economics 64 2017: 353-362

We explore four fundamental channels of mandate compliance available under current U.S. biofuel policy: increased ethanol blending through E10 or E85, increased biodiesel blending, and a reduction in the overall compliance base. Simulation results highlight the interplay and varying importance of these channels at increasing blend mandate levels. In addition, we establish how RIN prices are formed: The value of a RIN in equilibrium is shown to reflect the marginal cost of compensating the blender for employing one additional ethanol-equivalent unit of biofuel. This contrasts with existing research equating the price of RINs to the gap between ethanol supply and demand evaluated at the mandate level. We demonstrate the importance of this distinction in case of binding demand side infrastructure constraints such as the ethanol blend wall: as percentage blend mandates increase, the market for low-ethanol blends may contract in order to reduce the overall compliance base. This has important implications for implied ethanol demand in the economy.


Distinguishing Common Ratio Preferences from Common Ratio Effects Using Paired Valuation Tasks (with Kirby Nielsen, Ted O'Donoghue, Jason Somerville and Charles Sprenger) [pdf] [Online Appendix

Conditionally Accepted at American Economic Review

Without strong assumptions about how noise manifests in choices, we can infer little from existing empirical observations of the common ratio effect (CRE) about whether there exists an underlying common ratio preference (CRP). We propose to solve this inferential challenge using paired valuations, which yield valid inference under common assumptions. Using this approach in an online experiment with 900 participants, we find no evidence of a systematic CRP. To reconcile our findings with existing evidence, we present the same participants with paired choice tasks and demonstrate how noise can generate a CRE even for individuals without an associated CRP.

Taking a Load Off: Experimental Evidence of Preferences for Control with an Application to Residential Electricity Demand [pdf]

The rising share of renewable generation has led to an increased focus on demand-side mechanisms to balance the electric grid. For example, Direct Load Control (DLC) contracts allow utilities to curtail the electricity use of participating households at times of system stress. I use a novel experimental design to show that intrinsic preferences for control can significantly impact the rewards required to encourage consumers to participate in such contracts. In particular, I test for the existence, magnitude, and attributes of a control premium in a lab environment which mimics basic features of the DLC context. I find that participants on average exhibit a control premium of 9-32% above the instrumental value of the decision, which responds to both the probability and stakes of ceding control. There is limited evidence for the existence of an endowment effect with respect to control. Participants' stated motivations underlying their decisions are consistent with an inflation of (perceived) option value that cannot be explained by probability weighting.