Resumen de Three essays in applied economics
- español
En esta disertación, primero, estudio los efectos de Internet en los resultados individuales y regionales. Segundo, describo el espionaje político de el servicio secreto de Alemania del Este, la Stasi. En el capítulo uno, vinculo la disminución de la concentración de las distribuciones de votos en Alemania con los cambios individuales de las interacciones cara a cara frente a las virtuales. Muestro que las regiones donde los votantes participan en más interacciones a través de Internet exhiben distribuciones de votos menos concentradas. En el segundo capítulo, investigo si Internet influye en la integración de los inmigrantes en Alemania. Encuentro efectos positivos sobre las probabilidades de empleo, de hablar alemán y la frecuencia de actividades de voluntariado. En el capítulo tres, Albrecht Glitz y yo descubrimos el énfasis temático del espionaje de la Stasi. Además, proporcionamos evidencia de que la Stasi recopiló información sistemáticamente sobre las negociaciones entre ambos estados alemanes.
- English
In the first two chapters of this dissertation, I investigate how the shift from face-to-face towards virtual interactions, facilitated by the introduction of broadband internet, shapes individual and regional socio-economic and political outcomes. The basic intuition which underlies both chapters is the fact that individual decisions are prone to peer effects which lead to different outcomes for face-to-face and virtual interactions.
In chapter one, I show how the change in individuals’ communication patterns can explain the decline of the federal vote share distribution that is observable for Germany since the mid-1970s. Intuitively, real-life social ties tend to emerge relatively more frequently between individuals with different political opinions than it is the case for virtual social ties. Therefore, peer effects in face-to-face interactions have the potential to align individuals’ political opinions and thus concentrate the vote share distribution. Virtual social ties formed over the internet, on the other hand, are likely to emerge between voters of the same parties and thus cannot have any effect on vote share distributions.
I also build on the notion that research on mass media and peer effects suggests that the internet has significant effects on immigrant integration. The empirical studies investigating this relationship, however, focus only on a narrow set of integration dimensions and suffer from endogeneity. I try to fill these gaps with the paper presented in chapter two.
In the last chapter of this dissertation, Albrecht Glitz and I analyze a dataset on the political espionage activities of the so-called Stasi, the former East German secret service. On top of a comprehensive description, we also provide two examples of how to exploit the respective data for quantitative research, an approach that has not been implemented by many researchers in the field.
In chapter one, to show how virtual interactions affect aggregate political outcomes, I simulate a game with peer effects in voting for individuals placed on a network. Two parameters determine the concentration of the aggregate vote share distribution in the network. The overall number of social ties and the probability with which ties emerge between voters with similar partisan affiliation. Using this intuition, I link the decline in the concentration of German federal election vote shares since the mid-1970s to two key trends in the society. A decline of face-to-face and an increase of virtual interactions which take place relatively more often between individuals with similar partisan affiliation. To corroborate the assumptions of my model, for a sample of movers, I show that destination region vote shares significantly influence partisan affiliation. I provide evidence that the relationship is caused by peer effects and not by regional sorting. Using clubs and associations as a proxy for face-to-face interactions and the location of district courts in which they must be registered as an instrument, I confirm the model's prediction that regions where voters have fewer social ties exhibit less concentrated vote share distributions. Finally, by implementing a well-established instrumental variable approach for the number of households in a region connected to the broadband network, I show that more virtual interactions further contribute to the decline in the concentration of the vote share distribution.
In chapter two, I study a panel of foreign-born individuals in Germany and assess whether broadband internet exposure influences a comprehensive set of integration indicators. To identify potential effects, I regress individual changes in integration outcomes on changes in individual and household level broadband exposure measures. Furthermore, I use the previously mentioned instrumental variable to predict technical broadband availability at the municipality level. Since the instrument is not sufficiently strong to implement a two-stage least squares regression, I opt for a placebo-type approach instead: although technically not available in their municipality of residence, some households still bought a broadband access for their homes. I exploit this fact and show that individual and household level measures of broadband exposure are positively related to migrants' employment probabilities, their probability to usually speak German at home and the frequency of performing voluntary activities. However, in line with a causal interpretation, these effects are only significant for individuals living in municipalities where broadband internet is technically available. They are not significant for those individuals who bought a broadband access but live in regions where broadband internet is technically not available. There is some evidence for heterogeneity. The effect of broadband exposure on employment possibilities decreases with a migrant's age. Effects on German language use and contacts to Germans become smaller when immigrants use the internet to communicate with friends and family from their countries of origin.
In the last chapter, together with Albrecht Glitz, I provide an overview of the political espionage activities of the East German secret service, the so-called Stasi. The basis of our analysis are meta data on politically relevant material that was sent to the Stasi by its spies in the West. These data also contain information about the reports that were compiled on the basis of the incoming materials and sent to high-ranking politicians and friendly intelligence agencies. We begin by giving a basic description of the form, relevance, and content of the respective materials. We then describe the spies and recipients in terms of the number and relevance of materials provided\slash received. In two analyses, we then show how to exploit the meta data, especially keywords used as content labels, for further research. First, we provide evidence that the Stasi targeted negotiations between high-ranking East and West German representatives on a cultural agreement and a significant loan. Second, we fit a Latent Dirichlet Allocation (LDA) model to uncover the thematic emphasis of the Stasi's political espionage and how it evolved over time.
