Essays in applied microeconomics

Rusche, Felix

PDF
Rusche2025Thesis.pdf - Veröffentlichte Version
Download (26MB)

URN:	urn:nbn:de:bsz:180-madoc-699712
Dokumenttyp:	Dissertation
Erscheinungsjahr:	2025
Ort der Veröffentlichung:	Mannheim
Hochschule:	Universität Mannheim
Gutachter:	Ciccone, Antonio
Datum der mündl. Prüfung:	2025
Sprache der Veröffentlichung:	Englisch
Einrichtung:	Außerfakultäre Einrichtungen > GESS - CDSE (VWL) Fakultät für Rechtswissenschaft und Volkswirtschaftslehre > Makroökonomie (Ciccone 2013-)
Lizenz:	Creative Commons Namensnennung 4.0 International (CC BY 4.0)
Fachgebiet:	330 Wirtschaft
Fachklassifikation:	JEL: O12, J13, J16, J18, J71, J15, C93, J46, D85, L82, G10,
Freie Schlagwörter (Englisch):	mass media , policy , women empowerment , spatial econometrics , education , discrimination , job networks , labor markets , field experiment , media bias , financial markets
Abstract:	The media have been widely recognized for their dual capacity to contribute to both tragedy and large-scale social change. For example, mass media have been involved in events ranging from the Rwandan genocide to changes in gender roles in India and Brazil (Jensen and Oster, 2009; La Ferrara, Chong, and Duryea, 2012; Yanagizawa- Drott, 2014). Given their role in disseminating information, shaping attitudes, and influencing behavior, an important question arises: can policymakers harness this power to affect socioeconomic outcomes? This may be of particular interest in poor countries as the media offer a cheap and scalable way to reach large populations. In addition, evidence on the effects of media on gender roles underscores their potential as an instrument for social change. Yet, to comprehend the media's real world impacts, it is important to understand how journalists portray the world and, ultimately, decide what to report on. The significance of media in shaping socioeconomic outcomes is further underscored by the growing prominence of social media and online networks in economic, personal, and professional interactions. The digitization of these interactions not only allows for the study of media effects, but also provides researchers with unprecedented access to behavioral data that were once difficult to observe. For example, professional networking - long a critical yet elusive aspect of labor markets - can now be directly observed via platforms such as LinkedIn. The platforms further provide researchers with a controlled environment to cleanly study such phenomena using experiments. Not least, advances in statistical learning and econometric techniques offer powerful tools to mine these vast, high-dimensional, and unstructured data sets, opening new avenues for understanding the dynamics that underlie economic and social behavior. The three independent chapters of this thesis broadly engage with these questions. The first chapter studies the use of community media as a policy instrument by examining the effects of India's community radio policy on women's empowerment. The second chapter leverages the central role of LinkedIn in professional networking as an empirical setting to gain insight into discrimination against Black individuals in the process of building and using job networks. The third chapter focuses on how journalists decide what news to report on, specifically exploring whether news coverage disproportionately emphasizes negative events and why. As such, this thesis investigates the role of the media in shaping socioeconomic outcomes, exploits its growing importance to answer economic questions, and offers insight into how the media operates. In addition, it seeks to deepen our understanding of socioeconomic disparities by examining both the origins of these inequalities and the mechanisms through which they can be alleviated. To speak to these questions, the chapters adopt novel empirical settings, data sources, and econometric methods, including the development of an econometric method for causal identification, a multistage online experiment, and the use of statistical learning techniques for analyzing unstructured data such as image, text, spatial, and audio data. Below, I briefly summarize each chapter. Chapter 1: “Broadcasting Change: India’s Community Radio Policy and Women’s Empowerment” is motivated by the the fact that in poor countries, the interaction of early marriage, early motherhood, and low educational attainment disempowers women and limits their life opportunities. Even as countries grow richer, gender inequality is often sustained by social norms, thereby limiting welfare gains from women’s empowerment. This chapter investigates the use of media as a cheap and scalable policy to empower women. In 2006, India enacted a community radio policy that grants radio licenses to NGOs and educational institutions with the aim to foster local development. I collect original data on the content and coverage areas of all 250+ radio stations. I uncover women’s empowerment as a key theme through topic modeling and GPT-based analyses of radio show recordings. For identification, I exploit topography-driven variation in radio access and develop a novel econometric approach to deal with randomly displaced geolocated household data. The results show that women exposed to radio gain additional education and are more likely to obtain a secondary degree. In line with increased education, exposure reduces child marriages and fertility of young women while increasing their likelihood to exhibit autonomy in household decisions. The findings demonstrate that community media can effectively address gender inequality. Chapter 2: “LinkedOut? A Field Experiment on Discrimination in Job Network Formation” is motivated by the fact that around half of all jobs in the US are found via professional networks (Topa, 2011). At the same time, underrepresented groups tend to have smaller networks with fewer high-quality connections (Fernandez and Fernandez- Mateo, 2006). To understand to what extent this is driven by discrimination, the chapter assesses the impact of discrimination on Black individuals’ job networks across the U.S. using a two-stage field experiment with 400+ fictitious LinkedIn profiles. In the first stage, race is varied via A.I.-generated images only. The results show that Black profiles’ connection requests are 13% less likely to be accepted. Based on detailed information from users’ CVs, widespread discrimination is found across social groups. In the second stage, Black and White profiles are exogenously endowed with the same networks. Connected users are then asked for career advice. The results show no evidence of direct discrimination in information provision. However, when taking into account differences in the composition and size of networks, Black profiles receive substantially fewer replies. Overall, the findings suggest that gatekeeping is a key driver of Black-White disparities. Finally, Chapter 3: “Reporting Big News, Missing the Big Picture? Stock Market Performance in the Media” asks which events journalists choose to report on. More specifically, it studies whether news reporting disproportionately focuses the negative and why. This is studied in the context of news reporting on the stock market. It starts with a stylized fact: Between 2017 and 2024, the main national stock market indices rose in the US and the five largest European economies. However, the average daily performance of all six indices turns from positive to negative when weighted by daily media coverage. A case in point is the average daily performance of Germany’s DAX index on days it was reported on the country's most-watched nightly news. While the DAX increased by more than 4 index points per day over the period, the index dropped by more than 10 points on days it was reported - news was bad news. On days the DAX wasn't covered on the nightly news, the index rose by around 10 points - no news was good news. About half of the worse daily performance when the DAX was covered is accounted for by a greater focus on negative news. The other half stems from a novel big news bias: a greater focus on large index changes, whether positive or negative, combined with a negative skew in the daily performance of the index. The big news bias extends to other national stock market indices.