2026 Theses Doctoral

Essays in Policy and Behavior

Jeong, Hayeon

In this dissertation, I analyze the interaction between policy and behavior—how policy shapes behavior and how behavior, in turn, reshapes policy outcomes (sometimes undermining them, sometimes amplifying them). My first chapter focuses on the relationship between policy effectiveness and policy intensity through demand response channel. Many policies in practice become gradually more intensive over time in the hope of becoming more effective. While stricter policies may enhance compliance, they may also encourage loophole exploiting behaviors that offset intended policy benefits. If the increase in loophole exploitation outweighs the increase in compliance, overall policy effectiveness can actually decrease when a policy is intensified. I exploit the gradual intensification of Seoul’s vehicle control policy—an increasingly common policy instrument shaping pollution exposures in cities globally.

In my theoretical framework, I model individual decisions to comply, exploit loopholes, or violate the policy outright, and assess how these responses vary with policy stringency. Empirical results support the theoretical prediction that higher policy intensity can induce greater loophole exploitation. I find loophole exploiting behaviors—driving outside regulated hours—increased with policy tightening. Across multiple tightening events, the most stringent phase led to a 19% rise in loophole exploitation, the next to 13%, and the least stringent to 7%. I find a deterioration in air quality (PM10, PM2.5) by 8 to 10% during non-crackdown hours due to loophole exploitation. The resulting net health costs underscore the potential unintended consequences of intensified regulation. I show that health costs arising from loophole exploitation can be easily overlooked without careful evaluation. Accounting for these demand responses is therefore essential to accurately estimate the net health impact of policy intensification.

My second chapter analyze the impact of vehicle weight and smartphone penetration on pedestrian fatality and suggest the necessity of policy that address this problem. This is a joint project with Gernot Wagner and Stuti Saria. 1 Pedestrian deaths in the U.S. decreased steadily until 2010, after which they began to rise. This research investigates the causes behind this recent increase in pedestrian fatalities—a trend unique to the U.S.—and interprets it through the lens of pedestrian–driver inequality.

We hypothesize that the growing popularity of smartphones and heavy vehicles, such as SUVs, since 2010 has contributed to more frequent traffic accidents and more severe crashes, leading to the observed rise in pedestrian deaths. We decompose the relationship between heavy vehicles, smartphones, and non-occupant fatalities into two components: associations with crash frequency and, conditional on involvement in a traffic accident, associations with pedestrian fatality risk. We find that increased smartphone penetration is associated with a 0.02 to 0.05 increase in average crash frequency.

Furthermore, an additional 1,000 pounds of vehicle weight is associated with the risk of immediate fatality for a non-occupant by 12% in the event of a traffic accident. Driver distraction due to smartphone use also elevates this risk; conversely, a smartphone network outage reduces a non-occupant’s probability of immediate death by 3.5%. We emphasize the need for laws and regulations that protect non-occupants marginalized by current U.S. transportation and automobile policies. Specifically, we propose introducing incentives for the production of lighter electric vehicles (EVs), and enforcing stricter regulations on smartphone use while driving.

My third chapter is a behavioral and experimental project which looks at the impact of policy design on giving behavior. This is a joint project with Mark Dean, Jeffrey Guo, and Zhi Hao Lim. We seek to study whether giving the current “dictator" control over all future generations induces them to be more fair than if they had control over only themself. We suggest two “multigenerational” versions of the basic game: sequential dictator game (SEQ) and simultaneous dictator game (SIM), both of which are characterised by a sequence of players who represent the fictitious generations in an intergenerational allocation problem. In the SEQ, a decision is made in each round where the current dictator (generation) allocates the endowment (that is passed down from the previous generation) between himself and all future generations. This contrasts with the SIM where only the initial dictator (generation) gets to make a decision by simultaneously allocating the initial endowment between himself and every other future generation. This simple paradigm allows to investigate the effect of different settings on the decision process of resource allocation.

We found that subjects assigned to SEQ game keeps the endowment about 10% more compared to those assigned to SIM through between-subject design experiment. Additionally, we intend to measure correlation between the difference in giving behavior and idiosyncratic characteristics including risk attitude, belief in other players, and basic sociodemographic information, through within-subject design experiment. The results of the study will give new insights about generosity and fairness in absence of reciprocity, and provide insights about possible strategies to increase fairness in intergenerational coordination problems.

Taken together, these chapters highlight a common theme: policy effectiveness depends critically on how individuals and markets respond to the incentives and constraints that policies create. The first chapter shows how intensifying regulation can induce avoidance behavior that undermines its intended effects. The second chapter illustrates how, in settings where protections for non-occupants are limited, decentralized responses—such as the adoption of heavier vehicles and increased smartphone use—can generate negative externalities and exacerbate inequality between vehicle occupants and non-occupants. The third chapter demonstrates that differences in policy design can shape individual giving behavior, particularly in intergenerational contexts. Viewed together, these findings underscore that policy is not implemented in a vacuum: its success depends not only on its stated goals but also on the behavioral responses it induces. Designing effective policy therefore requires anticipating these responses and their impact on policy in return, and incorporating them into both the structure and evaluation of policy interventions.Env