Lina Kramer is a member of the Department of Political Economy at King's College London. Her research examines the intersection of artificial intelligence (AI) and social sciences. with a particular focus on how machine learning can support evidence‑based fiscal policy analysis and design. Her work evaluates both the opportunities and risks that AI technologies pose for modern fiscal governance.

She is funded by the UK Research and Innovation (UKRI) Economic and Social Research Council (ESRC) through the London Interdisciplinary Social Science Doctoral Training Partnership studentship award.

She holds an M.Sc. in Economics from the University of Cologne and a B.A. in Public Management and Governance from Zeppelin University. Prior to her doctoral studies, she worked for six years as a consultant supporting the digital transformation of the German public sector and federal administration.

Research Interests

• Fiscal policy, taxation, and government revenue
• Applications of machine learning in public policy
• AI-driven automation and its social, economic, and distributional impacts
• The dynamic interaction between economic inequality and redistributive policies
• Public opinion and political behaviour
• Climate policy and public support for climate-related fiscal measures
• Quantitative and computational methods

Doctoral research

Lina’s doctoral research investigates how machine learning can enhance fiscal policymaking while safeguarding democratic accountability. Her thesis consists of three components:

1. A systematic review of machine learning in taxation research, analysing 217 studies using methods ranging from PRISMA to NLP to trace the evolution of ML applications in tax policy, forecasting, administration, and enforcement.
2. Improving government revenue data quality and insights using machine learning, developing and evaluating ML-based imputation techniques for the Government Revenue Dataset (GRD) and using the resulting full dataset (1980–2022, 192 countries) to examine long‑term global trends in revenue levels and structures.
3. A survey experiment on AI chatbots and fiscal climate policy opinions, assessing whether recommendations attributed to ChatGPT influence public support for fuel taxation and climate investment in Germany and the UK.

In addition to the thesis chapters, Lina has also explored the use of multi‑agent reinforcement learning (MARL) for tax policy design. This project investigated whether MARL could simulate and optimise tax systems by modelling interactions between economic actors under constraints such as growth, inequality, and welfare. While the project was not included in the final thesis due to methodological challenges and time constraints, it reflects Lina’s broader interest in computational approaches to fiscal governance and the potential of AI‑driven models for policy experimentation.

Supervisors

Professor Konstantinos Matakos and Dr Maria Cotofan.