Harnessing Big Data for Advanced Risk Classification in Insurance Markets

• Applicant: Dr. Niklas Häusle
• Subject Area: Accounting and Finance ; Operations Management and Computer Science for Business Administration
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 563955509

Big data offers new opportunities to assess risks with greater precision, reduce adverse selection, and achieve socially important goals like more effective risk management. Building on this foundation, this research project explores innovative approaches to segmentation and prevention in insurance markets. Alongside improving efficiency in insurance markets, the project also focuses on providing policy recommendations for welfare-enhancing regulation. In doing so, the research bridges the fields of insurance science, actuarial science, and big data analytics. The first work package investigates whether digital footprints, such as behavioral indicators (e.g., the time required to complete an application form), have predictive power in the context of motor insurance. For this purpose, I link footprint data with claims data and compare their predictive power with traditional risk variables such as occupation and age. The second work package focuses on the aspect of risk prevention and its detection through big data. High-resolution aerial imagery is processed with machine learning algorithms in two steps: First, I use object recognition to transform the image data into identifiable objects. Then, I analyze how these identified objects, particularly preventable risks such as nearby trees or missing roof shingles, influence hail risk. This preventive approach complements the risk classification perspective of the first work package by aiming to actively reduce risks. The third work package addresses the regulatory challenges of big data risk classification. The goal is to examine how policymakers can regulate the segmentation enabled by the vast array of variables from big data, with a focus on the welfare implications of various regulatory strategies. This work package places particular emphasis on translating the findings into understandable recommendations for lawmakers and the public. It links the themes of the first two work packages and considers their implementation in regulatory contexts that are necessary to enable a welfare-maximizing use of granular data sources. Thus, the first work package directly contributes to improving efficiency in insurance markets. The second work package expands this perspective by incorporating the prevention aspect, enabling insurers to reward preventive measures with financial incentives and thereby promoting a system that reduces losses and keeps premiums affordable. Building upon the insights gained in the first two work packages and considering their potential implementation within an appropriate regulatory framework, the third work package can ultimately develop policy recommendations for welfare-enhancing regulation.

• DFG Programme: Research Grants