This page collects selected academic reports, presentations, and data-related projects from my coursework and independent work. The focus is on applied economic analysis, market design, behavioural economics, and reproducible research.
Market Design & Contract Theory
Comparison of Discriminant Analysis and Random Forests in Classification — Bachelor's Thesis
Type: Bachelor's Thesis | Format: Thesis (PDF) & Reproducible R Code | Field: Statistical Classification, Machine Learning, Applied Statistics
A bachelor's thesis comparing two classification approaches on the same applied task: classical linear and quadratic discriminant analysis (LDA/QDA) as a parametric, model-based benchmark, versus random forests as a nonparametric, ensemble-based machine-learning method. Implemented entirely in R, the analysis evaluates both methods in terms of predictive accuracy, interpretability, computational cost, and robustness across a real-world classification problem. The repository contains the full reproducible R code, all generated figures, and the complete thesis as a PDF — making the workflow easy to inspect, re-run, and extend. The work is released under a CC BY-NC-ND 4.0 licence.
Key questions addressed:
- How do linear and quadratic discriminant analysis perform on a real classification problem relative to a tuned random-forest baseline?
- What are the trade-offs between parametric interpretability and nonparametric flexibility as model complexity and sample size vary?
- How do the two approaches differ in handling class imbalance, feature interactions, and decision boundaries?
- To what extent does ensemble averaging in random forests improve stability and generalisation compared to a single discriminant rule?
Voluntary Disclosure and Personalized Pricing — A Market Design Analysis
Type: Market Design Analysis | Format: Report & Presentation | Field: Market Design, Information Economics, Industrial Organization
This project examines how firms' incentives to voluntarily disclose product or quality information interact with their ability to price discriminate based on consumer data. Drawing on information economics and market design theory, the analysis explores when voluntary disclosure emerges as an equilibrium strategy, how personalized pricing alters firms' disclosure incentives, and what the resulting welfare implications are for consumers and market efficiency. The written report combines a structured literature review with formal economic reasoning, while the accompanying presentation condenses the key mechanisms and findings into a concise, visually supported summary for an academic audience.
Key questions addressed:
- Under what conditions do firms disclose information voluntarily absent regulation?
- How does access to personalized consumer data change disclosure incentives?
- What are the efficiency and distributional consequences of personalized pricing enabled by disclosed information?
Binary Payment Schemes: Moral Hazard and Loss Aversion — Academic literature review presentation
Type: Literature Review | Format: Academic Presentation | Field: Contract Theory, Behavioural Economics, Market Design
A literature review presentation examining Herweg, Müller, and Weinschenk (2010, American Economic Review), which explains a longstanding puzzle in contract theory: why real-world incentive schemes are typically simple binary bonus contracts, even though classical principal–agent theory (Holmström, 1979) predicts fully contingent, continuous wage schedules. The paper resolves this by introducing expectation-based loss aversion (following Kőszegi & Rabin, 2006/2007) into the standard moral hazard framework, modelling how agents evaluate wages relative to their anticipated distribution rather than a fixed reference point. The presentation traces the theoretical mechanism, the resulting trade-off between incentive provision and loss-aversion costs, and the empirical evidence for expectation-based reference points in effort and labour supply decisions.
Key questions addressed:
- Why do simple, binary bonus schemes dominate in practice despite theoretical predictions of complex contracts?
- How does expectation-based loss aversion change the optimal design of incentive contracts?
- What trade-off exists between finer performance-based incentives and the psychological cost of loss aversion as contract complexity increases?
- What empirical support exists for expectation-based reference points in labour markets?
Behavioural & Climate Economics
Misperceived Social Norms and Climate Action — Presentation
Type: Literature Review | Format: Academic Presentation | Field: Behavioural Economics, Climate Policy, Social Norms
A presentation examining Bergquist & Nilsson (2024, Review of Economics and Statistics)¹, which studies whether Americans misperceive the prevalence of climate-friendly behaviour and support for climate policy among their peers — and whether correcting these misperceptions changes real, incentivised behaviour. Using a two-wave survey experiment (n = 2,000 baseline; n = 6,000 information experiment), the study finds that respondents substantially underestimate both the share of Americans trying to fight global warming (43% perceived vs. 71% actual) and the share supporting climate action (65% perceived vs. 87% actual). An information treatment correcting these beliefs increased incentivised donations to a climate charity (atmosfair) by $3.20–$5.00, with effects concentrated among respondents who initially underestimated climate-friendly norms.
Key questions addressed:
- Which behavioural factors predict willingness to act against climate change?
- Do Americans underestimate public support for and engagement in climate action?
- Can correcting these misperceptions causally increase climate-related donations and policy support?
- Are treatment effects heterogeneous across individuals with different prior beliefs?
Misperceived Social Norms and Climate Action — Report
Type: Policy Analysis | Format: Written Report | Field: Behavioural Economics, Climate Policy, Social Norms, EU Policy
A seminar report analysing Andre, Boneva, Chopra & Falk (2024, The Review of Economics and Statistics), which shows that Americans systematically underestimate both the prevalence of climate-friendly behaviour and public support for climate action — a phenomenon of "pluralistic ignorance." Correcting these misperceptions through low-cost information treatments causally increases incentivised donations and stated policy support, with the strongest effects among climate sceptics. Building on this evidence, the report develops an original policy argument for the EU context: as instruments like the revised Energy Performance of Buildings Directive (EPBD) and ETS2 begin to impose direct, visible costs on households, public acceptance — not instrument design — becomes the binding constraint on decarbonisation. The report argues that norm-correction campaigns can dissolve the "phantom majority" of perceived opposition, complementing rather than replacing carbon pricing, and situates this argument within the broader literature on social influence, second-order beliefs, and the limits of nudges and rebate-based policy communication.
Key questions addressed:
- Why do individuals systematically misperceive social norms around climate action, and what does this cost in terms of behaviour and policy support?
- How does correcting these misperceptions differ mechanistically from conventional nudges?
- Can norm-correction interventions increase the political feasibility of binding EU climate instruments such as ETS2 and the EPBD?
- What are the structural limits of belief-based interventions in the face of liquidity constraints, split incentives, and polarised media ecosystems?
Research & Technical Focus
- Topics: market design, behavioural economics, climate economics, information economics
- Methods: theoretical analysis, literature review, empirical reasoning, data visualisation
- Tools: LaTeX, GitHub, R, HTML