Metadaten







Promotionsordnung


PromO07

Kumulative Dissertation


ja

Titel


Essays in Applied Microeconometrics

Titel (englisch)



Autor/Autorin


Steinmayr, Andreas

Unter Mitarbeit von


Felfe, Christina; Huber, Martin; Lechner, Michael; Mahmoud, Toman Omar; Rapoport, Hillel; Trebesch, Christoph

Geburtsdatum


08.03.1983

Geburtsort


Rohrbach/Österreich

Matrikelnummer


09616475

Schlagwörter (GND)


Sport; Humankapital; Matching; Migration

DDC (Dewey Decimal Classification)


Sozialwissenschaften - 300

Freie Stichwörter (deutsch)


Mikroökonometrie; Kinderentwicklung; Stichprobenselektion

Freie Stichwörter (englisch)


Sports; child development; radius matching; sample selection; principal stratification

Kurzfassung


Ziel dieser Dissertation ist es, die Ursachen und Konsequenzen von menschlichem Verhalten mittels empirischer Forschungsmethoden zu untersuchen. Obwohl die einzelnen Kapitel sehr unterschiedliche Themen behandeln, verbindet sie das Ziel, einen kausalen Mechanismus zu identifizieren oder entsprechende Methoden für andere empirische Arbeiten zu entwickeln bzw. bestehende Methoden zu verbessern.

Das erste Kapitel untersucht ob sportliche Aktivität von Kindern ihre kognitiven und nicht-kognitiven Fähigkeiten beeinflusst. Dieses Kapitel basiert auf zwei Datensätzen zu Kindern in Deutschland und verwendet semi-parametrische Schätzmethoden, um für die Selbstselektion in sportliche Aktivität zu kontrollieren. Die Ergebnisse zeigen, dass sportliche Aktivität die schulische Leistung und das Gruppenverhalten verbessert.

Das zweite Kapitel untersucht die Sensitivität eines Radiusmatching Schätzers in Bezug auf verschiedene Tuningparameter mittels einer empirischen Monte Carlo Simulation. Die Ergebnisse dieser Studie helfen Anwendern die passenden Werte für die einzelnen Parameter zu wählen. Im Rahmen dieses Papers wurde der Schätzer für die Softwarepakete GAUSS, STATA und R implementiert.

Das dritte Kapitel untersucht den Effekt von Emigration auf das Wahlverhalten von im Ursprungsland verbleibenden Haushaltsmitgliedern. Die dahinter liegende Hypothese ist, dass Migranten in den Zielländern über die dortigen politischen Institutionen lernen und dieses Wissen an im Ursprungsland verbleibende Angehörige weitergeben. Diese Studie basiert auf administrativen und Umfragedaten der ehemaligen Sowjetrepublik Moldawien. Die Ergebnisse zeigen, dass die Migration eines Individuums nach Westeuropa dazu führt, dass die zurückbleibenden Haushaltsmitglieder eher liberale, pro-westliche Parteien wählen. Migration nach Russland hat keinen derartigen Effekt.

Das vierte Kapitel beschäftigt sich mit der Identifikation des kausalen Effekts der Migration eines Individuums auf die zurückbleibenden Haushaltsmitglieder. Die Schwierigkeit für die Identifikation dieses
Effekts ist die gleichzeitige Existenz von mehreren Selektionsproblemen. Die bestehende Literatur adressiert üblicherweise die nicht zufällige Selektion von Haushalten in Migration. Das darüber hinausgehende Selektionsproblem, dass manche Haushalte nur einen Migranten senden, während in anderen Fällen der ganze Haushalt emigriert, wurde bisher nicht systematisch behandelt. Die vorliegende Untersuchung zeigt, welcher Bias durch dieses zweite Selektionsproblem entstehen kann und wie der kausale Effekt partiell identifiziert werden kann. Mittels der vorgeschlagenen Methode wird der kausale Effekt von Migration auf den Schulbesuch von Kindern in Mexiko analysiert. Die Ergebnisse zeigen, dass die Migration eines Erwachsenen einen negativen Effekt auf den Schulbesuch von Jungen hat. Der Effekt auf Mädchen ist hingegen nicht eindeutig.

Kurzfassung (englisch)


The aim of this thesis is to understand the nature and consequences of human behavior by using empirical research methods. At the core of each chapter is the aim to reveal causal mechanisms - the question whether a particular mechanism can explain an observed correlation, for example that families of Moldovan migrants, who emigrated to Western countries tend to vote for liberal parties or that sporty children do better in school. The inherent problem is that a causal interpretation of correlations is often difficult as they might also be the result of selection. The proposed papers have in common that they either try to overcome the problem of endogeneity and to reveal the forces at work in different fields of the economy and society or that they address methodological problems related to the identification of causal effects.

The first paper investigates the effects of physical activity on children's development. While the importance of cognitive and non-cognitive skills for success in life is widely acknowledged, their formation is still only partly understood. In particular, the role of curricular activities for the formation of health and human capital has been widely studied. Yet, the role of extra-curricular activities has received only little attention. This study uses two different datasets from Germany to analyze the effect of participation in a sports club on various measures of health and cognitive and non-cognitive skills. We use flexible semi-parametric estimation methods with a specific way to use the panel dimension of the data to address selection into sports. We find positive and robust effects of sports on children's school performance and peer relations, but not on health. We also find that increased physical activities replaces passive leisure activities like watching TV. This substitution may partially explain the effect.

The empirical approach used in the first paper requires comparing children who do sports with children who do not do sports, but who have otherwise very similar characteristics. It is relatively easy to find similar or even identical individuals as long as the number of characteristics is small, the number of potential comparison individuals is large, and there is sufficient overlap in the distributions of treated and control individuals. However, the problem becomes much more complex once the number of characteristics grows - a problem that is commonly known as the curse of dimensionality. One way to overcome this problem is propensity score matching. Instead of matching on all characteristics, one can also match on a single index that expresses the propensity to be in the treatment group. However, this technique requires to choose several tuning parameters. The second paper uses a simulation design that is based on empirical data, to investigate the finite sample behavior of radius matching with respect to various tuning parameters. In particular we explore the behavior of distance-weighted radius matching with bias adjustment. The results are intended to help the practitioner to choose suitable values of these parameters when using this method. We furthermore implement this estimator as radiusmatch command in the software packages GAUSS, STATA and R.

The remaining two chapters focus on the effects of international migration. The third paper investigates the effects of emigration on voting behavior of those who stay behind. The classical perspective on the relationship between emigration and politics is shaped by Hirschman's classical theory of Exit, Voice, and Loyality. This theory proposes that individuals dissatisfied with the political/institutional system can either push for change (voice) or emigrate (exit). This theory suggests that the emigration of dissatisfied individuals helps regimes to stay in power. We argue that emigration creates political externalities beyond this exit effect. Migrants are exposed to different institutions in the destination country. This exposure changes their perception about the institutions in the home country and they might become more critical about the institutions in their home country, even if they did not emigrate for political reasons in the first place. However, this effect varies with the political regime of migrants' destination countries. We furthermore argue that migrants transfer the newly acquired information and insights to their peers who stayed behind. Using community and individual-level data from Moldova, we analyze the relationship between emigration to different destinations and electoral preferences of those who stay behind. For identification, we exploit the fact that emigration from Moldova is directed to both more democratic Western Europe and less democratic Russia. Our key finding is that emigration to the West reduces electoral support for the Communist Party at home. Quantitatively, the effects are large enough to explain the fall of the long-ruling Communist government.

The fourth paper investigates identification of the effects of one household member on the educational attainment of children left behind. What makes identification challenging is the presence of more than one selection problem. The literature usually looks at the effects of emigration of one household member on the remaining members and addresses the non-random selection of the household into migration. However, while in some households only one member migrates, in other households some or all household members migrate. In the latter case, the household will usually not be included in cross-sectional data at all. Gibson et al. (2010, 2011a) refer to this problem as the double-selection problem. We tackle this problem by modeling the behavior of the household members using principal stratification. This allows identifying bounds on the effects of migration on school attendance of children who stay behind in Mexico. The results suggest that adult migration reduces school attendance rates of boys while the direction of the effect is ambiguous for girls.

Universität


Universität St.Gallen

Referent/Referentin


Lechner, Michael (Prof. Dr.)

Korreferent/Korreferentin


Bauer, Thomas (Prof.)

Erweitertes Diss. Komitee


Kirchgässner, Gebhard (Prof.)

Fachgebiet


Economics and Finance (PEF)

Sprache


ENG

Promotionstermin (dd.mm.yyyy)


17.02.2014

Erstellungsjahr (yyyy)


2013

Dokumentart


Dissertation

Format


PDF

Dissertationsnummer


4137

Quelle



PDF-File


dis4137.pdf

Dokumentverknüpfung


Link zu diesem Dokument







letzte Änderung: 03/18/2014 - Allgemeine rechtliche Informationen - Datenschutz [ Nach oben ]