Detecting insurance fraud using supervised and unsupervised machine learning

Fraud is a significant issue for insurance companies, generating much interest in machine learning solutions. Although supervised learning for insurance fraud detection has long been a research focus, unsupervised learning has rarely been studied in this context, and there remains insufficient evidence to guide the choice between these branches of machine learning for insurance fraud detection. Accordingly, this study evaluates supervised and unsupervised learning using proprietary insurance claim data. Furthermore, we conduct a field experiment in cooperation with an insurance company to investigate the performance of each approach in terms of identifying new fraudulent claims. We derive several important findings. Unsupervised learning, especially isolation forests, can successfully detect insurance fraud. Supervised learning also performs strongly, despite few labeled fraud cases. Interestingly, unsupervised and supervised learning detect new fraudulent claims based on different input information. Therefore, for implementation, we suggest understanding supervised and unsupervised methods as complements rather than substitutes.     

The dynamic effects of monsoon rainfall shocks on agricultural yield,wages, and food prices in India*

This paper shows that, first, the effects of monsoon rainfall shocks on agricultural yield in India are highly asymmetric: yield falls strongly after droughts, whereas excessive rainfall has only little effects. Second, our key novel finding is that the short-lived yield loss after a widespread drought elicits a persistent decline (increase) in wages (food prices), which lasts for up to five years. Third, affiliation to the same National Sample Survey region (and thus to the same state) seems to be a key determinant of internal migration, whereas distance appears more relevant for food arbitrage trading.     

Vox Populi,Vox Dei? Tacit collusion in politics

We study competition between political parties in repeated elections with probabilistic voting. This model entails multiple equilibria, and we focus on cases where political collusion occurs. When parties hold different opinions on some policy, they may take different policy positions that do not coincide with the median voter's preferred policy platform. In contrast, when parties have a mutual understanding on a particular policy, their policy positions may converge (on some dimension) but not to the median voter's preferred policy. That is to say, parties can tacitly collude with one another, despite political competition. Collusion may collapse, for instance, after the entry of a new political party. This model rationalizes patterns in survey data from Sweden, where politicians on different sides of the political spectrum take different positions on economic policy but similar positions on refugee intake—diverging from the average voter's position, but only until the entry of a populist party.     

Regulation of Location-Specific Externalities from Small-Scale Polluters

Environmental and Resource Economics - Emission damages caused by small-scale polluters such as farms, vehicles, homes and small businesses are often location-specific and such polluters are often...     

Open innovation ecosystem organizing from a process view: a longitudinal study in the making of an innovation hub

This paper uses process theory as a theoretical lens to analyze AstraZeneca's enactment of an open innovation initiative with the purpose of strengthening the firm's surrounding innovation ecosystem. Based on empirical data collected over 7 years, we develop a process model of open innovation enactment and explain how the initiative gradually transformed while maintaining its guiding principles, which were set from the start. In applying a process perspective, we highlight open innovation initiatives as dynamic and evolutionary – but not deterministic – developments. As such, we provide a comprehensive and more nuanced understanding of not only what open innovation is but also how it becomes. This study also contributes to the innovation ecosystem literature by theorizing how firms orchestrate innovation ecosystems through open innovation initiatives over time.     

OPENING THE FLOODGATES: PARTIAL AND GENERAL EQUILIBRIUM ADJUSTMENTS TO LABOR IMMIGRATION

We investigate the impact of a large immigration shock on occupational wages. We develop a general equilibrium model where individuals sort into occupations and confront testable hypotheses with data. To identify the effect of the labor supply shock, we introduce a novel instrument that exploits that immigrants systematically sort into different occupations than natives. We study the immigration wave to Norway after the Eastern enlargement and find that immigration led to lower relative occupational wages. A quantification of the general equilibrium shows welfare effects of immigration close to zero for natives, but negative effects for the pre-existing population of immigrants.