Dissertations submitted to the Hertie School
Refine
Year of publication
- 2024 (3) (remove)
Document Type
- Doctoral Thesis (3)
Language
- English (3)
Has Fulltext
- yes (3)
Is part of the Bibliography
- no (3) (remove)
03/2024
This dissertation comprises three essays. Although their titles make clear their self-contained nature, all essays share a common motivation: to combine economic theory with modern tools of causal inference to understand applied problems.
The first chapter studies honesty when the costs to behave honestly change. Using field data from a snack delivery company that employs an honesty payment system, the chapter presents an event study to analyze how price increases affect pay rates. The honesty payment system expects customers to pay a listed price for each consumed snack. This allows me to measure honesty with the pay rate that compares consumption to payments. The results, which draw on deliveries from several thousand firms, show that price increases that make honest behavior more costly cause more cheating. Price increases of 15% trigger a fall in pay rates of 11%.
The second chapter studies a recent legal reform in Germany, which aims to lower commission rates of real estate agents by raising the cost salience of sellers. I find that the reform has backfired and real estate agents have exploited the transition to increase their commission rates. The findings document that in some regions real estate agents increase their commission by up to 2 percentage points, adding over e6,000 in transaction cost to the average home sale. As explicit collusion is unlikely in this setting, I argue that this arbitrary increase points to seller ignorance instead. To verify if and why sellers fail to induce price competition, I run a pre-registered survey experiment with 1,062 real estate agents. Although commission rates should be negotiated independently for each sale, the survey confirms that 85% of sellers do not attempt to negotiate lower commission rates. The randomized experimental questions suggest that real estate agents may cater to the low willingness of sellers to negotiate by providing misleading reference commission rates and shrouding the economic incidence for sellers.
The third chapter is co-authored with Christian Traxler and Carsten Burhop and examines the causal effect of beer on crime based on unique panel data from Germany between 1882 and 1913. Using exogenous variation in the yield of spring barley, a key input in beer production, we identify a quantitatively and qualitatively significant effect of beer on violent crimes. This effect is mainly driven by a positive effect on assaults, where our findings suggest that a 1% increase in beer production raises assault rates by the same percentage. These findings are corroborated using a complementary empirical design that exploits a brewing tax reform in Prussia.
02/2024
This thesis is a comprehensive study of trust. The study rests on the argument that all forms of trust stem from the same universal concept. Once trust has been defined at its core, the concept can be specified within a wide variety of contexts, such as trust in your friends and family, in your fellow citizens, or in different kinds of institutions within your country. Although trust is an increasingly popular topic within the social sciences, a lack of consensus on the true meaning of the concept remains. One the one hand, this can be attributed to diverging conceptualisations in a rich and multidisciplinary conceptual literature. On the other, to empirical studies where quantifiable measures do not spring from a clear conceptual account, but rather take survey responses directly at face value. Based on this fundamental issue, the overarching goal of my thesis is to bridge the gap between the conceptual and empirical literature.
I achieve this goal with three contributions. Firstly, I conduct a conceptual analysis of trust, which results in a unifying conceptual framework. The framework acknowledges and combines different conceptual understandings from political science, psychology and sociology, thereby representing a move towards clearer consensus on the meaning of trust within the conceptual literature. Secondly, I develop a new measurement model for trust, which is directly rooted in the conceptual framework. The model reflects an alternative approach to empirical research, where quantitative measures of trust are derived from a conceptual argument on the appropriate structure and dimensionality. Thirdly, I identify a typology of citizens who trust in fundamentally different ways. The typology acts as a complementary perspective on the meaning of trust, studied from the viewpoint of citizens in a heterogeneous population.
As a whole, I argue that the thesis represents a new approach to trust research, where conceptual and empirical methods are closely intertwined. In order to advance our understanding of this fundamentally important concept, it is crucial to bring the two strands of the literature closer together.
01/2024
In the past decade, social media has revolutionized how governments create public value through online communication and collaboration with citizens. Despite this, academic literature still lacks a comprehensive understanding of how public value is disseminated via social media. This research delves into the process of public value creation on social media in New Zealand, a leading country in leveraging social media for government-to-citizen communication, particularly during crises.
The study seeks to answer two questions about creating or eroding public value over social media. The first question investigates the socio-demographic factors correlating with citizens' varying perceptions of public value from social media adoption. The findings suggest that citizens with similar public value perceptions share common socio-demographic characteristics, such as age, gender, education, trust in government, platform usage, and following government accounts. The study also reveals differences in public value perceptions based on the platform type, with platforms like Twitter garnering more positive public value perceptions than Facebook.
The second question examines how using social media in government can enhance or diminish public value perceptions among citizens. The research constructs a causal model where specific social media practices during peaceful and crisis times, along with the engagement of external stakeholders, lead to increased perceptions of public value creation. The study also uncovers two causal mechanisms for the erosion of public value. In the first model, government reactions to opposing opinions on social media, such as hiding or removing comments, result in low public value perceptions. The second model shows that over-posting or posting irrelevant information can reduce citizens’ trust in the government, limit government posts' reach, and lower public value perception.
This research encourages future studies to replicate the experiment across various platforms and countries to better understand how public value can be created or eroded over social media.