http://hdl.handle.net/1893/90 Electronic theses of Economics Students. Fri, 25 Sep 2026 09:00:35 GMT 2026-09-25T09:00:35Z http://hdl.handle.net/1893/38246 Title: Geopolitical risk: examining the incidence and impact of geopolitical risk uncertainty in Africa Author(s): Atanga, Isaac Anamsakya Abstract: This thesis develops a locally grounded framework for measuring and analysing geopolitical risk (GPR) in African economies. It argues that geopolitical risk is context-specific and cannot be adequately captured by global indices alone, particularly in settings where domestic insecurity, political instability, local media narratives, and regional security dynamics shape economic expectations. Using domestic newspaper sources, locally relevant keywords, and supervised machine-learning validation, the thesis constructs new GPR indices and applies them to three empirical questions: whether local GPR predicts macroeconomic outcomes, whether multiple uncertainty sources explain emerging-market volatility, and how geopolitical risk is transmitted across African economies. Chapter Two shows that Nigeria’s Local Geopolitical Risk index contains predictive information that is largely missed by global benchmarks. Increases in local GPR are associated with naira depreciation, changes in investment dynamics, and weaker consumer spending, even after controlling for global oil- price movements. Chapter Three demonstrates that South African equity volatility is shaped by domestic geopolitical risk, global policy uncertainty, and macro-financial conditions within a GARCH–MIDAS framework, confirming that emerging-market volatility reflects multiple uncertainty channels operating at different frequencies. Chapter Four extends the analysis regionally by constructing comparable local GPR indices for eight African economies and estimating spillovers using Diebold–Yilmaz and Baruník–Krˇehlíkconnectednessmethods. Theresultsrevealsubstantial,asymmetric,andhorizon- dependent regional spillovers. Gravity-model evidence further shows that formal trade linkages are not the dominant transmission mechanism once endogeneity is addressed. Instead, spillovers are strongly shaped by Sahel–Lake Chad security exposure and by recipient-country absorptive capacity. Overall, the thesis contributes to geopolitical-risk measurement, African macro-finance, and regional- risk analysis. It shows that locally measured geopolitical uncertainty affects exchange rates, capital flows, consumption, volatility, and regional spillovers. The findings support the use of local GPR indices in macroeconomic surveillance, reserve management, fiscal-risk assessment, stress testing, and AfCFTA- related regional early-warning systems. Sun, 12 Oct 2025 00:00:00 GMT http://hdl.handle.net/1893/38246 2025-10-12T00:00:00Z http://hdl.handle.net/1893/38135 Title: Advancing and Applying Modern Counterfactual Analysis in Public Economics and Policy Analysis Author(s): Belshe, Lindon Scott Abstract: One popular approach to policy analysis emphasises the construction of counterfactuals as a means of isolating the causal effects of some intervention or treatment – often a policy interventions or broader social changes. These alternative trajectories allow researchers to estimate both short- and long-term impacts of treatments or events across diverse contexts. Yet, because the value of counterfactual analysis depends on the credibility of its underlying construction, growing attention has turned to improving how counterfactuals are modeled and applied. Better counterfactual construction not only strengthens causal inference but also broadens the scope of counterfactual analysis itself; it transforms it from a tool limited to well-defined, date-specific experiments into one capable of evaluating complex, historical, and systemic policy processes. Traditional empirical methods may encounter difficulties in addressing “what-if” scenarios, especially when interventions are challenging to isolate or change gradually over time. In recent years, alternative approaches to counterfactual construction have been developed to enhance the methodological rigor and applicability of quasi-experimental policy analysis. For example, the synthetic control method and machine learning–based simulation models provide options for estimating counterfactual outcomes with increased reliability and adaptability. The unifying objective of this dissertation is to advance these modern approaches to counterfactual construction for improved policy analysis in public economics, demonstrating how they can yield more credible insights across different domains and can be applied to important but challenging questions. Through three distinct but thematically linked studies, the research showcases how cutting-edge counterfactual techniques enhance our ability to measure policy impacts and inform evidence-based decision-making. Together, the three papers that follow demonstrate the application of these approaches across distinct yet complementary areas of public economics. The first examines long-run health outcomes through the lens of Scotland’s life expectancy divergence, using synthetic control methods to identify when and how its trajectory began to deviate from comparable countries. The second extends this framework to fiscal policy, evaluating the economic and mortality effects of Ireland’s corporate tax under cutting through a similar but more rigid counterfactual design. The third shifts from retrospective evaluation to prospective simulation, applying machine learning–based valuation and redistribution models to explore revenue-neutral reforms to England’s Council Tax system. Collectively, these studies illustrate how refined counterfactual construction can illuminate complex economic and social processes that lie beyond the reach of conventional empirical tools. 1. A Comparative Analysis of the Scottish Divergence in Life Expectancy at Birth The first paper, A Comparative Analysis of the Scottish Divergence in Life Expectancy at Birth, applies synthetic control analytics within a historical and demographic context. The study aims to explore one of the most enduring puzzles in public health economics, which is Scotland’s persistent lag in life expectancy relative to other Western nations. In the paper, a “synthetic Scotland” is constructed from a weighted combination of comparable countries and covariates to estimate what Scotland’s life expectancy trajectory might have been in the absence of its observed divergence. This counterfactual framework allows for an exploration and reassessment of when Scotland’s mortality path began to deviate, tracing long-run trends across well over a century of data. The findings reveal that Scotland’s life expectancy deficit likely emerged in two waves, with the first occurring shortly after World War I. This is far earlier than the previously assumed mid-century period The initial divergence set in motion a slower rate of health improvement that persisted through the 20th century. The analysis also confirms a secondary divergence during the 1970s and 1980s, coinciding with economic and social restructuring across the UK. The second wave is more in-line with prior theories. Beyond identifying the timing of these breaks, the study disaggregates results by age and sex, finding that younger working-age populations have long accounted for the greatest burden of excess mortality (and thus shorter life expectancies). Additionally, while male life expectancy has remained comparatively stable in its deficit relative to other countries, the female gap has widened since the 1970s. By applying the synthetic control framework to contexts without discrete policy interventions, it is demonstrated that counterfactual analysis can serve as a robust exploratory tool to identify latent structural changes beyond simply assessing known treatments. This methodological advancement enhances our understanding of Scotland’s public health trajectory and illustrates how constructing credible long-term counterfactuals can provide valuable insights into historical developments influencing current health inequalities. 2. Death and Taxes: Does Corporate Tax Undercutting Result in Lower Economic Growth and Lower Mortality? The second paper, Death and Taxes: Does Corporate Tax Undercutting Result in Lower Economic Growth and Lower Mortality?, extends the counterfactual framework from demographic trends to fiscal tax policy, examining the Republic of Ireland’s dramatic corporate tax reductions as a natural experiment in economic and social development. Between 1997 and 2003, Ireland reduced its statutory corporate tax rate from 36% to 12.5%, creating one of the most competitive tax environments in the OECD. Using the SCM, a “synthetic Ireland” is constructed, which represents the country’s potential trajectory devoid of these tax reductions. This framework enables a comparative analysis of both macroeconomic and public health outcomes, focusing on foreign direct investment (FDI) per capita, gross national income (GNI) adjusted for purchasing power parity, and mortality rates. The analysis indicates that Ireland’s corporate tax policies, which involved significant reductions compared to marginal rate changes elsewhere, were associated with increases in FDI and GNI, as well as enhanced competitiveness and economic activity relative to previous trends in tax rates and trade liberalization. However, it is difficult to isolate the impact of corporate tax measures from broader social and economic developments in the late 1990s and early 2000s. The period known as the Celtic Tiger included multiple factors such as EU integration, educational improvements, and technological growth, which make direct attribution complex. Still, the evidence suggests that substantial tax reductions played a role in Ireland’s rapid economic expansion. The analysis also explores associated health outcomes, with results that are more uncertain but nonetheless suggestive. Improvements in mortality relative to the synthetic counterfactual appear to follow the period of accelerated economic growth, though these effects are subject to lags and noise in the data. The evidence tentatively points to a link between economic gains and better population health, and Ireland’s current international standing in mortality rates supports this possibility. Taken together, these findings highlight both the power and the nuance of counterfactual policy analysis. While complete causal separation is rarely attainable in such dynamic historical contexts, the construction of robust counterfactuals allows researchers to draw meaningful and credible inferences about how large-scale fiscal shifts can influence broader social and health outcomes. By integrating mortality considerations into fiscal policy evaluation, this study expands counterfactual analysis to encompass welfare effects beyond traditional economic metrics. This approach constitutes an innovative synthesis of public finance and health economics within a cohesive empirical framework. Additionally, the paper illustrates that the synthetic control method is effective not only for assessing the direct causal impact of policies but also for identifying secondary effects, such as changes in life expectancy or mortality rates. Ultimately, the research underscores the capacity of counterfactual analysis to connect microeconomic and macroeconomic perspectives, thereby enhancing our understanding of how structural policy decisions shape outcomes in both economic and human development. 3. Reforming Council Tax in England: Empirical Evidence and Reform Options for a Fairer Property Tax System The third paper, Reforming Council Tax in England: Empirical Evidence and Reform Options for a Fairer Property Tax System, turns from retrospective policy evaluation to forward-looking simulation. It develops a data-driven framework to assess how England’s property tax system, which is currently anchored to 1991 valuations and relies on a rigid and regressive banding structure, could be modernised to improve both equity and efficiency. In the study, a comprehensive property-level dataset is constructed using administrative records on property characteristics, historical transactions, and tax liabilities. A geographically weighted Random Forest–based Automated Valuation Model (AVM) is then constructed to estimate contemporary property values. Using these updated valuations, simple revaluation benchmarks and thirteen counterfactual tax system simulations are constructed. These range from modestly adjusting the existing bands to significantly reweighting tax liabilities to even more ambitious structural changes, such as value-based mill levies. Each scenario is evaluated under a revenue-neutral constraint, ensuring comparability to the current system while isolating the effects of valuation updates and structural design choices. The findings reveal a deeply misaligned tax base, where more than half of dwellings are estimated to fall into inappropriate bands when rank-ordered nationally. This far exceeds other estimates that focus on more local revaluations. This finding implies that high-value properties are systematically undertaxed and lower-value homes are overtaxed. Thus, a simple statistical revaluation alone corrects much of this imbalance, improving horizontal equity by aligning liabilities more closely with market values. However, additional gains arise from introducing finer banding or proportional rate structures, which enhance progressivity and reduce effective tax disparities across regions and income groups. Beyond the empirical findings, this study contributes methodologically by combining machine learning–based valuation with counterfactual policy simulation. It establishes a framework for evaluating property tax reforms at a national level that is both flexible and reproducible. The model allows policymakers to examine potential outcomes of various design choices prior to implementation, providing quantifiable comparisons regarding simplicity, fairness, and fiscal stability. While the paper does not recommend a single system, it presents evidence that periodic revaluation and incremental structural adjustments could impact fairness in local taxation. Overall, the study demonstrates how computational methods can be applied to policy evaluation, facilitating the analysis of reform concepts through testable policy scenarios. Collectively, these studies show that counterfactual methods in public economics are advancing in both technique and application. As new data sources and computational tools expand the precision and scalability of such models, the potential for counterfactual measurement to inform real-world policy decisions will only grow. Future innovations, whether through integrating causal inference techniques with machine learning or developing more dynamic forms of synthetic control, will further enhance the capacity of researchers and policymakers to evaluate interventions in complex, interdependent systems. In this sense, the work presented here not only demonstrates current applications but also points toward a rapidly maturing frontier in the quantitative evaluation of public policy. Sat, 01 Nov 2025 00:00:00 GMT http://hdl.handle.net/1893/38135 2025-11-01T00:00:00Z http://hdl.handle.net/1893/37518 Title: Five essays on valuing non-market ecosystem services of natural resources: insights from environmental attitude, perceptions, and choice behaviours Author(s): Jeon, Chulhyun Abstract: This dissertation, Five Essays on Valuing Non-market Ecosystem Services of Natural Resources: Insights from Environmental Attitude, Perceptions, and Choice Behaviors, investigates how behavioral drivers and contextual factors shape the valuation of forest ecosystem services. By combining meta-analysis, contingent valuation, discrete choice experiments, and hybrid choice models, it offers both methodological innovations and policy-relevant insights into the economics of conservation and restoration. The first essay tests the hypothesis that ignoring seasonal variation biases economic valuation of ecosystem services. Using a meta-dataset of 476 Korean observations, robust mixed-effects models reveal that willingness to pay (WTP) rises by 67% in autumn but falls by 18% and 65% in summer and winter, respectively. The results confirm that temporal heterogeneity systematically distorts annualized values if seasonality is omitted. Economically, this underscores the need for dynamic valuation models that align with the cyclical supply of ecosystem services. The second essay examines whether the design of payment vehicles systematically alters conservation values. In the case of the Gotjawal forests of Jeju, contingent valuation surveys with 1,100 residents reveal that entrance fees elicit the highest WTP and policy acceptance, while tax-based mechanisms face strong resistance. This supports the hypothesis that valuation outcomes are not neutral to institutional design, highlighting the economic principle that the incidence and framing of payments critically shape demand for public goods. The third essay explores the hypothesis that public preferences for restoration are heterogeneous and segmented. A discrete choice experiment with 1,021 citizens, analyzed through principal component analysis and latent class choice modeling, identifies five preference segments with divergent WTP patterns. Acceptance declines steeply at higher costs, validating theoretical expectations of downward-sloping demand for public environmental goods. The findings highlight that efficient and feasible restoration policy requires aligning marginal benefits with socially acceptable cost thresholds. The fourth essay advances the hypothesis that latent environmental attitudes exert significant and systematic influence on restoration choices. Using a hybrid choice modeling framework that integrates mixed logit with latent class analysis, the study finds that environmental attitudes explain substantial variation in WTP and improve model fitness. This confirms the economic significance of incorporating unobserved behavioral drivers into preference models, bridging revealed and latent components of choice and extending the frontier of stated preference methods. The fifth essay investigates whether respondents’ choice behavior exhibits systematic deviations from fully rational attendance. Advanced econometric analysis uncovers selective attribute non-attendance and heterogeneous decision types, from deterministic to probabilistic choosers. These results challenge the assumption of fully compensatory preferences, providing empirical evidence that bounded-rationality and cognitive heuristics shape environmental decision-making. From an economic perspective, this stresses the importance of behavioral realism in welfare measurement and policy appraisal. Taken together, these essays demonstrate that the economic valuation of non-market ecosystem services is deeply conditioned by seasonality, payment design, attitudinal heterogeneity, and behavioral processes. The dissertation strengthens the methodological foundation of environmental economics by refining stated preference approaches and integrating hybrid choice models. At the same time, it delivers actionable insights for conservation finance: that socially legitimate and economically efficient policies must reflect behavioral drivers and contextual realities rather than abstract market analogies. Mon, 01 Jan 2024 00:00:00 GMT http://hdl.handle.net/1893/37518 2024-01-01T00:00:00Z http://hdl.handle.net/1893/37390 Title: Advancing Choice Modelling for Time-Critical Decision-Making: Integrating Machine Learning to Improve Predictive Accuracy and Interpretability Author(s): Murali Parthasarathy, Kavya Abstract: Aim: This thesis examines how time pressure affects decision-making in dynamic driving scenarios, integrating traditional discrete choice modeling with modern ma- chine learning approaches to improve predictive accuracy, interpretability, and un- derstanding of behavioral adaptations over time. Methods: An online experiment with 514 participants was conducted, where each participant made 10 simulated driving decisions under varying time pres- sures—from unlimited time to severe 5-second constraints. In each trial, participants were asked to choose whether to stay in their current lane or change lanes, making lane choice the primary dependent variable. Study 1 used mixed-effects logistic re- gression to explore how time constraints influenced optimal lane-change decisions, while also accounting for individual differences such as risk perception, education level, and gender. In Phase 2 of the study, these insights informed the develop- ment of a feature- engineered Long Short-Term Memory (LSTM) model, trained to detect temporal patterns in lane-choice behavior. Phase 3 extended this by training an LSTM directly on raw, sequential input data-removing preprocessing constraints to understand the intricate latent behavior which was not captured by traditional models. Results: Time pressure had a non-linear impact on decision performance, with par- ticipants performing worst under moderate constraints (10–30 seconds), supporting the hypothesis that cognitive conflict peaks when there is insufficient time to delib- erate but too much to act on intuition. Conversely, performance improved under both low and extreme time constraints, indicating shifts in cognitive strategy (e.g., deliberative vs. heuristic modes). In Phase 2, we transitioned from inference to pre- diction, using a feature-engineered LSTM model to address the limitations of static regression models. Although it improved predictive accuracy its performance was still constrained by the reliance on hand-crafted features. In Phase 3, a raw-sequence LSTM model captured temporal dependencies and adaptive learning more effec- tively, achieving 95.6% test accuracy and 97.5% AUC-ROC. SHAP analysis and Par- tial Dependence Plots (PDPs) further explained the non-linear feature interactions and adaptive cue weighting, enhancing model transparency. Conclusion: This multi-method thesis demonstrates that decisions under time pres- sure are not merely degraded, but adaptive and context-sensitive. The integration of behavioral theory with interpretable deep learning offers new pathways for build- ing neuroadaptive decision-support systems in high-stakes domains such as au- tonomous transport, space operations, and cognitive augmentation technologies. Mon, 30 Sep 2024 00:00:00 GMT http://hdl.handle.net/1893/37390 2024-09-30T00:00:00Z