The project consists of two parts:
Part 1: We want to derive the optimal way of taxing a good that has a substantiative externality (for example airplane travel). We were inspired by the discussion about the Green New deal in the US which was criticized for including a strong redistributive component along with rules for emission reduction. What are the optimal policies for taxing such a good under different objectives? First results indicate that Pigouvian taxes always favor the rich relative the optimal utilitarian mechanism. Interestingly, non-market mechanisms (such as a fixed number of flights for each person) can be optimal. A major part of the RA project would be to take these results to the concrete application of airplane travel and quantify optimal taxes and consumption caps.
Part 2: Recent advancement in artificial intelligence and machine learning techniques greatly enhances the ability to predict individuals' potential outcome (e.g., performance, productivity, willingness to pay, risk factors, default probabilities). Nonetheless, it is also well-documented that the advancement of prediction powers comes at the expense of fairness, leading to a trade-off between prediction accuracy and fairness. This project studies the scope of information that can be uncovered under the constraint that prediction outcomes must be fair. The pre-doctoral fellow will read and summarized the theoretical literature on information design and the economics of AI, understand institutional details of algorithm design in various industry, write simple and demonstrative prediction algorithms, and conduct exemplary computation using actual data under the guidance of the faculty sponsor.
REQUISITE SKILLS AND QUALIFICATIONS:
Successful candidates are expected to have a solid mathematical background (calculus, real analysis, linear algebra, probability theory, proof-based math courses), are able to understand and summarize microeconomic theory research, and are proficient in basic programming (Python, Matlab). Applicants must have a bachelor's degree or master's degree.
SPECIAL APPLICATION INSTRUCTIONS:
Finalists will be asked to perform sample programming exercises and write summaries for several assigned papers.