This thesis seeks to develop a deeper understanding of the market characteristics that influence and enhance the profitability of pricing strategies employed by platform owners in multi-sided markets. It will also identify the optimal set of strategies for maximizing profits based on these underlying market characteristics. Along this, the study presents a model that encompasses three key participants in software platform markets: consumers, complementors, and platform owners. Consumers engage with applications offered by complementors to derive utility, paying a price in return. Complementors develop and offer applications on the platform, leveraging the platform’s technology infrastructure. Platform owners provide the platform and collect revenue through variable transaction commission fees or fixed royalties from complementors. The model operates in two stages: platform owners first select pricing strategies, followed by consumers and complementors making participation choices, considering single-homing consumers and multi-homing complementors.

The study identifies optimal pricing strategies, particularly favoring variable fees over royalties, in platforms where more than 50% of complementors are participating on. We show that sustainable oligopolies in platform markets depends on the existence of stable second-stage equilibria, where multiple platforms can secure sufficient consumer and complementor participation to cover costs and generate returns. The number and nature of these equilibria influence market outcomes, with scenarios of multiple stable equilibria often favoring monopoly tendencies with one platform holding the majority of the single homing consumers. We further argue that platform owners prefer single second-stage equilibria over constellations with multiple equilibria, and that even central second-stage equilibria are less favorable from a profit maximization perspective under risk neutrality. We provide a map of recommended strategies for incumbents and entrants subject to the general market conditions, with complementor heterogeneity and consumer benefits as the main parameters.

Application of the model to the real-world case of mobile phone applications demonstrates some discrepancies to empirical data, which can be overcome by adjusting our parameter set. This example showcases that real world competition on two platforms favors the existence of a single equilibrium on the second-stage game, which differs from literature on one platform cases.

Seasonal adjustment aims to remove predictable seasonal fluctuations and other calendar-related variations from a time series. In the context of economics Hylleberg (1992) defines seasonality  as ``[...] the systematic, although not necessarily regular, intra-year movement caused by the changes of the weather, the calendar, and timing of decisions, directly or indirectly through the production and consumption decisions made by the agents of the economy''.

The removal of seasonal and calendar effects allows analysts, economists, and statisticians to more clearly identify the business cycle and other non-seasonal effects, such as strikes or crises. By adjusting for seasonality, it becomes easier to compare data across different time periods and to make more timely economic assessments and policy decisions. The usual alternative, year-on-year growth-rates, show underlying developments with a time lag of about half a year (cf. Dagum and Mazzi 2018). Moreover, these growth rates may be skewed by  differences in the calendar constellation, for instance a discrepancy in the number of working days between the periods under comparison. To address these issues, seasonal adjustment techniques are employed to provide a more immediate and undistorted view of economic developments.

This dissertation adds to the literature on seasonal adjustment and time series econometrics by introducing contributions in two primary areas. Historically, the development and application of seasonal adjustment methods have predominantly focused on time series data with lower frequencies, such as monthly and quarterly observations. This work, however, develops and analyses methods to seasonally adjust higher frequency time series. Most prominently, a procedure is developed to adjust daily time series. 

The second area of contribution lies in the analysis and enhancement of diagnostic tools that are instrumental in the seasonal adjustment process for both lower and higher frequency time series. This includes the examination and refinement of seasonality tests, which are pivotal in detecting seasonal patterns within the data, as well as unit root tests, which are relevant in many areas of econometrics, and essential for the RegARIMA model based estimation of calendar effects. Collectively, these methodological advancements and diagnostic improvements facilitate more accurate seasonal adjustments.

An individual’s health over their lifetime is shaped by various factors, including access to healthcare, the quality of healthcare provided, as well as social and socioeconomic factors such as the physical and social environment, income, and education. The degree to which people adopt healthy lifestyles and take preventive measures can be considered another important determinant of health. This dissertation is composed of 4 empirical papers which each seek to explore the causal effects of a policy interventions or external shock in Germany on individuals’ health outcomes or relevant determinants of health.

The first paper examines the impact of local disease outbreaks on preventive health behavior, specifically vaccination rates. Using health insurance claims data, the study examines how regional measles outbreaks in Germany affect vaccination behaviors for children and adults. By exploiting temporal and geographic variation in outbreak. patterns, the paper identifies a causal relationship between local outbreaks and an increase in timely vaccinations among children, as well as catch-up vaccinations in adults. Notably, the effect is localized to the regions experi- encing outbreaks, and the behavioral changes are short-lived, suggesting that perceived disease risk – rather than a rational, long-term reassessment of the benefits of vaccination – drives the observed increase in preventive behavior.

The second and third papers focus on a recent education reform in Germany, known as

the G8 reform, which reduced the number of years in the upper secondary school track while maintaining the same curriculum. This reform, which intensified learning by compressing the number of instructional years, provides a unique opportunity to assess its unintended consequences on students’ mental and physical health. The second paper uses survey data. From the German Socio-Economic Panel Study (SOEP) to examine self-reported mental and physical health outcomes. The results reveal that the reform significantly decreased students’ mental health, particularly affecting vitality and emotional balance. However, it did not show a measurable impact on physical health.

The third paper extends this analysis by using healthcare claims data to investigate the incidence of diagnosed mental and physical health conditions among students. Contrary to the survey data findings, the analysis suggests that there is no significant increase in diagnosed mental disorders or stress-related physical conditions as a result of the reform. This apparent contradiction highlights the complexity of measuring health effects across different data sources and underscores the importance of using multiple methodologies to triangulate findings in policy evaluation research.

Finally, the fourth paper shifts focus to the healthcare system and explores the role primary care physicians as gatekeepers in the German healthcare system. In Germany, gatekeeping programs allow patients to access specialist services only upon referral from their primary care provider. Theoretically, such programs aim to improve coordination of care and control healthcare costs. Using healthcare claims data and a quasi-experimental approach, the paper examines whether participation in a physician-centered gatekeeping program leads to better health outcomes, higher quality of care, and changes in healthcare costs. Employing multiple identification strategies – including an instrumental variable approach and physician fixed effects. The

study finds that gatekeeping modestly improves care coordination and preventive care, while also leading to increased ambulatory costs. These findings contribute to the ongoing debate about the efficiency of gatekeeping programs in healthcare systems.

This dissertation explores the relationship between labor markets, public health policies, and human capital formation using historical data from Sweden. It comprises four empirical essays. The first analyzes a 1920 policy reducing working hours, showing significant long-term health benefits. The second and third chapters investigate the 1918–19 influenza pandemic, highlighting how labor disruptions affected mortality. The final chapter examines the impact of hospital births on long-term educational and health outcomes. 

This thesis contributes to the field of probabilistic forecasting in energy markets through novel methodological developments and empirical applications. The liberalization of energy markets over the past 25 years, combined with the transition towards carbon-neutral energy systems, has created increasingly complex and interconnected markets characterized by high uncertainty from renewable energy sources, demand fluctuations, and geopolitical risks. Traditional point forecasting methods are inadequate for capturing this uncertainty, necessitating probabilistic approaches that enable better risk management and decision-making.
The thesis makes several methodological contributions to the field of probabilistic online learning. First, we introduce CRPS learning, a novel aggregation method that combines probabilistic forecasts using flexible weight functions across quantiles rather than static weights. This approach accounts for varying forecast performance within different parts of the predictive distribution. Second, we extend this framework to multivariate settings, developing a method that exploits dependencies between combination weights of marginal distributions using penalized splines. Third, we propose a regularized probabilistic online learning algorithm that combines online coordinate descent with the framework of generalized additive models for location, scale, and shape (GAMLSS), enabling direct prediction of distributional parameters in high-dimensional settings.
The thesis also addresses specific forecasting challenges in energy markets through three empirical applications. We develop probabilistic forecasting models for day-ahead and month-ahead natural gas prices, accounting for heavy tails and asymmetric effects while considering the influence of European Emission Allowances (EUA) prices. We propose a multivariate VECM-Copula-GARCH model to jointly forecast EUA, natural gas, coal, and oil prices, capturing long-term equilibrium relations, conditional heteroscedasticity, and time-varying dependence structures through one-step estimation. Additionally, we explore the prediction of high-resolution electricity peak demand from low-resolution data using both statistical and machine learning approaches.
Complementing these methodological and empirical contributions, we address the computational challenges of implementing probabilistic forecasting methods by developing an R package for benchmarking C++ code. This tool provides overlapping and nested timers with OpenMP support and nanosecond-level precision, facilitating the development of efficient implementations for online learning algorithms.
The proposed methods are particularly well-suited for streaming data applications in energy markets, where models must update rapidly with new observations. Through simulation studies and empirical applications using real energy market data, we demonstrate the practical utility and superior performance of our approaches compared to established benchmarks. This work advances both the theoretical understanding of probabilistic forecasting and its practical implementation in energy market applications, contributing to more informed decision-making in an increasingly uncertain energy landscape.

Zeit, die mit Selbstmanagementaktivitäten bei chronischen Erkrankungen verbracht wird, Diese steht nicht mehr für alternative Tätigkeiten wie Arbeit oder Freizeitaktivitäten zur Verfügung. Das hat möglicherweise Einfluss auf die Bereitschaft von chronisch kranken Menschen, sich an Empfehlungen zu Selbstmanagementaktivitäten zu halten. Die Promotion bearbeitet am Beispiel des Diabetes mellitus vier Projekte: 1) Angesichts des limitierten Wissens über Patientenzeit, war es Ziel von Projekt 1, Zeitaufwand für Selbstmanagement bei Diabetes gesamt und für verschiedene Aktivitäten im Einzelnen zu untersuchen, zudem mit Patientenzeitaufwand assoziierte Faktoren. Genutzt wurden Daten einer populationsbasierten Studie. 2) Projekt 2 zielte darauf ab, Patientenzeitkosten zu analysieren, und baute unmittelbar auf Projekt 1 auf. 3) Angesichts des limitierten Wissens über die Frage, ob Menschen mit einer chronischen Erkrankung Zeitaufwand für Selbstmanagement in die Bewertung ihres Gesundheitszustands oder ihrer Lebensqualität einbringen, war es Ziel von Projekt 3 zu untersuchen, ob Zeitaufwand mit der Bewertung des Gesundheitszustands assoziiert ist. Dafür wurde die gleiche Population genutzt wie in Projekten 1 und 2. 4) Projekt 4 untersuchte mittels einer systematischen Literaturrecherche, ob und ggf. wie in gesundheitsökonomischen Evaluationen von Interventionen zur Behandlung und zur Prävention des Diabetes, die körperliche Aktivität als wesentlichen Bestandteil des Selbstmanagements beinhalteten, Patientenzeit für Bewegung explizit auf der Kostenseite einbezogen wurde. Projekt 1 ergab, dass mit fast 90% die Mehrzahl der Studienteilnehmenden angab, in der letzten Woche Zeit für Diabetes-Selbstmanagement aufgewendet zu haben. Insgesamt wurde mit rund 2,5 Stunden pro Woche ein substantieller Zeitaufwand angegeben. Diabetes-Komplikationen und Behandlungsarten waren erwartungsgemäß mit Zeitaufwand assoziiert, soziodemographische Faktoren jedoch nicht. Projekt 2 ergab Patientenzeitkosten von rund 2000 Euro pro Jahr. Sie lagen damit in der Größenordnung von indirekten Kosten und würden direkte medizinische Kosten sogar übersteigen. In Projekt 3 wurde in Modellen ohne Interaktion kein unabhängiger Zusammenhang zwischen Zeitaufwand und dem Gesundheitszustand gefunden. In Modellen mit Interaktion hatte der Zeitaufwand jedoch einen statistisch signifikanten Einfluss. Zusammenhänge scheinen demnach komplex und nicht linear zu sein. In Projekt 4 konnte nur eine begrenzte Zahl von Studien zur gesundheitsökonomischen Evaluation von Bewegungsinterventionen im Rahmen von Prävention und Management des Diabetes aus gesellschaftlicher Perspektive gefunden werden (n=23). Von diesen berücksichtigte rund die Hälfte der Studien Patientenzeiten (n=11). Weder in den Studien mit noch in den Studien ohne Berücksichtigung von Patientenzeiten wurde das Vorgehen erläutert oder diskutiert. Auffällig war die fehlende oder unzureichende Dokumentation der Methoden für die Erhebung und monetäre Bewertung des Zeitaufwands. Zusammenfassend wurde geschlussfolgert: 1) Der Zeitaufwand des Diabetes-Selbstmanagements ist substantiell und sollte bei der Betrachtung von Programmen und Interventionen in der Diabetesversorgung nicht ignoriert werden. 2) Ebenso sind Patientenzeitkosten substantiell, potentiell ähnlich hoch wie indirekte Kosten und sogar höher als direkte medizinische Kosten in dieser Patientengruppe, und sollten bei der Diskussion der ‘burden of disease’ des Diabetes berücksichtigt werden. 3) Der Impact von Zeitaufwand für Selbstmanagementaktivitäten auf die Bewertung des Gesundheitsstatus scheint – wenn denn existent – klein und komplexer als eine einfache lineare Assoziation. 4) Zeitaufwand für Selbstmanagementaktivitäten sollte in gesundheitsökonomischen Evaluationen berücksichtigt werden. Problematisch sind die Erhebung und die monetäre Bewertung des Zeitaufwands. Eine Standardisierung der Methoden ebenso wie der Dokumentation dieser Methoden ist

This thesis examines the impact of oil market shocks on the US economy. Chapter 1 offers an introduction to the topic and provides an overview of the remaining four chapters. 

Chapter 2 extends the existing understanding of the relation between oil prices and the US stock market in two ways: (i) by evaluating the effects of oil market shocks on the US stock market both, at an aggregate level and for all 49 US industry specific portfolios, and (ii) by scrutinizing the dynamic nature of this relation within a Structural Vector Autoregression (SVAR) model for a large set of rolling samples. The empirical results indicate that the effects of oil prices on the US stock market depend on the type and the timing of the shocks. An oil supply shock generally does not have a statistically measurable impact on stock market performance. Conversely, an aggregate demand shock has a positive effect on nearly all sectors, while an oilspecific demand shock has a negative, but statistically insignificant impact on stock returns for most industries. These results suggest that the importance of investors’ efforts to hedge against oil market shocks might be declining over time. In addition, the costs of carbon taxes and/or tradeable permits may be smaller than commonly thought if stock prices represent the net present value of profits.

Chapter 3 examines whether the alternative identification techniques for oil market shocks could be responsible for the mixed results in the oil-stock market literature. This study employs a Bayesian Structural Vector Autoregression (SVAR) to compare the implications of traditional identification approaches (SVAR with zero/sign restrictions) with those from the baseline model (Bayesian SVAR) for the case of the US. We find that the baseline model implies more plausible posterior price elasticities of oil supply and demand and a more profound effect of oil supply shocks on oil prices. Nonetheless, all models provide qualitatively similar conclusions for the effects of oil market shocks on the US stock market, with demand-side shocks playing a more important role than oil supply shocks. Overall, this study reveals that traditional identification schemes remain a good approximation in practice for the oil-stock market relationship.

Chapter 4 provides a novel perspective on the role of US oil production in both, the global oil market and the US economy. It employs a Bayesian Structural Vector Autoregression for the global oil market to identify distinct demand and supply oil market shocks. It differs from earlier studies by distinguishing explicitly between US and non-US oil production. In addition, a rolling-sample estimation approach is employed to investigate oil market shocks in a dynamic framework. The study finds considerable time variation in the effects of oil market shocks on the US economy. In particular, it identifies a sharp increase in the price elasticity of US oil supply after the shale oil revolution in the US. This increased price responsiveness of US oil production is partly responsible for the resilience of the US economy to global oil market shocks in recent years. This conclusion remains robust regardless of whether the global oil market shocks originate from the demand or supply side.

Finally, Chapter 5 summarizes the most important contributions of the thesis. It also tries to link the econometric methodologies and results in the individual chapters and aims to provide an overarching picture of the results in this thesis.

This thesis comprises four essays examining how institutions, economic policies, and culture shape gender inequality, demographic change, and investment in human capital. Reducing inequality and fostering human capital are key to achieving sustainable growth in developing countries.

Chapter 2 studies how local political institutions affected education policy in early twentieth-century Sweden. Municipalities transitioning from direct to representative democracy adopted school extensions earlier, suggesting that representative systems were more education-oriented. However, entrenched local elites often resisted reform, highlighting challenges of decentralization when elites dominate local agendas.

Chapter 3 explores the impact of cultural norms on gender inequality in Vietnam. Despite strong son preference, the rising sex ratio stems entirely from prenatal selection rather than girl neglect. Girls perform as well as or better than boys in health and education, challenging common assumptions about son preference and discrimination.

Chapter 4 analyzes how trade liberalization shaped women’s marriage and fertility decisions in Vietnam. The U.S.–Vietnam Bilateral Trade Agreement led young women to delay marriage and childbirth, with no significant effects on labor participation or sex ratios. These outcomes likely reflect structural changes and extended schooling for women.

Chapter 5 investigates how women’s economic empowerment affects domestic violence in rural Bangladesh. Women contributing economically to their households face significantly more verbal abuse, underscoring the need to recognize non-physical forms of domestic violence.

Overall, the thesis broadens understanding of how political structures, cultural practices, and socio-economic contexts interact to influence gender equality and human capital development, offering insights for policies promoting inclusive and sustainable growth.

Die vorliegende Dissertation untersucht die Integration von drei bislang weitgehend isoliert behandelten Themengebieten an der Schnittstelle zwischen Betriebswirtschaftslehre, Wirtschaftsinformatik und (Kern-)Informatik. Dabei handelt es sich um Projektmanagement und Wissensmanagement aus betriebswirtschaftlicher Perspektive sowie um Künstliche Intelligenz (KI) aus der spezifischen Perspektive von Ontologien und Case-based Reasoning, die vorwiegend in der Wirtschaftsinformatik und Informatik erforscht werden.

Im Fokus steht die computergestützte Aufbereitung von Wissen aus überwiegend natürlichsprachlichen Texten im Kontext öffentlicher Bauprojekte und Projektausschreibungen. Insbesondere wird die Generierung von Wissenskomponenten zur Erstellung und Verarbeitung computergestützter Projektbeschreibungen betrachtet. Diese sollen im Rahmen des Case-based Reasonings das betriebliche Projektmanagement unterstützen, insbesondere als Planungsgrundlagen für die Teilnahme an und die Ausführungsplanung von öffentlichen Bauprojekten.

Die Arbeit befasst sich mit drei zentralen Herausforderungen:

1. Die Extraktion von projektbezogenem Wissen aus schlecht strukturierten und überwiegend natürlichsprachlichen Dokumenten wie "Lessons Learned", "Debriefings" und "Project Reports". Für die betrachtete Domäne öffentlicher Bauprojekte betrifft dies insbesondere die formale Repräsentation ausschreibungsbezogenen Wissens zur computergestützten Weiterverarbeitung und Kombination mit Erfahrungswissen aus früheren Bauprojekten.
2. Die Anwendung des Text-to-Onto-Konzepts aus der traditionellen, symbolischen KI-Forschung zur semantischen Wissensverarbeitung. Dieses wird mit Case-based Reasoning kombiniert, um projektbezogenes Erfahrungswissen zu integrieren. Dieser Ansatz grenzt sich bewusst von rein syntaktischen, probabilistischen Verfahren der subsymbolischen KI-Forschung ab, die aktuell durch Deep Learning Networks dominiert wird.
3. Die Untersuchung und Evaluation geeigneter KI-Tools für die praktische Umsetzung des Text-to-Onto-Konzepts. Dabei werden Effektivitäts- und Effizienzaspekte hinsichtlich der Qualität der generierten Wissensrepräsentationen und des erforderlichen Ressourcenaufwands berücksichtigt. Zudem wird die Differenzierung zwischen KI-Tools und den darauf basierenden Case-based-Reasoning-Systemen betrachtet.

This thesis develops and applies advanced estimation techniques that combine multiple identification principles or leverage statistical evidence across observational units to enhance variable selection and inference for parametric models. Emphasis is placed on shrinkage estimation. The analysis focuses on longitudinal data, especially on distinguishing unit roots from mean-reverting time series.

Chapter 2 introduces information-enriched weighting of the regressor used for unit root inference in adaptive Lasso estimation of autoregressions. By incorporating additional evidence on the stochastic order of the data, the approach improves model selection performance. The asymptotic properties of this modified Lasso estimator are derived, with particular attention to the stochastic properties of the Lasso solution path.

The theoretical results of Chapter 2 underpin a novel, tuning-free significance testing principle based on the Lasso path, proposed in Chapter 3. This method bridges adaptively ℓ1-penalised model selection with unit root testing in autoregressions.

Chapter 4 addresses the challenges of heteroskedastic and autocorrelated empirical processes by proposing a sieve wild bootstrap algorithm that makes the inference procedures from Chapter 3 robust, ensuring accurate size under a general class of error processes.

Chapter 5 contributes to the panel unit root literature by introducing an autoregressive wild bootstrap algorithm for meta-analytical unit root tests, enabling robust inference under time-varying volatility and dynamic cross-dependence.

Chapter 6 applies Bayesian hierarchical models that leverage information from data-rich observational units to improve inference for units with sparse information via shrinkage.

The developed methods are applied to a range of empirical questions with economic, environmental, and societal relevance. These include modelling inflation rates, testing for trends in house price dynamics, analysing convergence in environmental indicators such as groundwater levels and CO₂ emissions, as well as estimating penalty conversion abilities in professional football.

Beim Management von Projekten ist die intelligente Wiederverwendung von Erfahrungswissen ein weit verbreitetes, aber wenig adressiertes Realproblem. Zur Lösung dieses Realproblems wird im gegenwärtigen State of the Art die integrative Nutzung der drei Themengebiete Projektmanagement, Wissensmanagement und Künstliche Intelligenz – in Form von Ontologien und Case-based Reasoning – diskutiert.

Die vorliegende Arbeit knüpft an den gegenwärtigen State of the Art an, indem sie die intelligente Wiederverwendung von Erfahrungswissen im Projektmanagement mittels der Kombination von Ontologien und Case-based Reasoning in den Vordergrund rückt. Durch die Verknüpfung von Ontologien und Case-based Reasoning wird eine inhaltliche, d. h. semantische Wissensverarbeitung ermöglicht. Zugleich wird jedoch durch die besondere Konzentration auf Adaptionsregeln ein Themenfeld fokussiert, das im gegenwärtigen State of the Art des ontologiegestützten Case-based Reasoning weitgehend unbeachtet bleibt.

Neben der Konzentration auf Adaptionsregeln wird eine neuartige Projektmanagementontologie vorgestellt, die umfassende sprachliche Ausdrucksmittel für ein ontologiegestütztes CBR-System, das zur Wiederverwendung von Erfahrungswissen im betrieblichen Projektmanagement dienen soll, spezifiziert und Charakteristiken von Commonsense-Ontologien, Domänen-Ontologien und Methodenontologien aufweist.

L<L