Year: 2026
Type: dissertation
Abstract: This dissertation develops models and analytical tools for studying the role of uncertainty in the adoption of carbon capture, utilization, and storage (CCUS) technology, and contributes new methodology to the field of quantile-parameterized distributions (QPDs). The first two chapters focus on CCUS policy. Chapter 2 develops a twostage stochastic programming model for optimal CCUS infrastructure network design and explores the effects of policy uncertainty and investor risk aversion on CCUS infrastructure development. Applied to a case study of the Texas-Louisiana Gulf Coast, we find that the current 45Q tax credit levels justify meaningful near-term CCUS investment regardless of future policy outcomes. Furthermore, we find that policy uncertainty is a more consequential barrier to CCUS deployment than investor risk aversion, reducing expected CO2 capture by at most 16% across the scenarios considered in our case study. Chapter 3 examines carbon capture (CC) subsidies from the government’s perspective using a Stackelberg game model with a social-welfaremaximizing government as the leader and a profit-maximizing firm as the follower. We analytically derive the optimal subsidy, establish conditions under which a CC subsidy may lead to increased net CO2 emissions, and show that government uncertainty about firms’ CC costs generally raises the optimal subsidy. We also conduct a case study of a coal-fired power plant and find an optimal subsidy value of approximately 83/tCo2, remarkably close to the current U.S. 45 Q level of 85/tCO2. The CCUS models in the first two chapters rely on simple distributional assumptions that may not reflect the limited information available in practice. The final two chapters address this limitation by developing new tools within the field of quantile-parameterized distributions, which offer flexible methods for constructing distributions from sparse quantile data. Chapter 4 introduces the QFlex distribution, a highly flexible and general QPD system designed to exactly match (or fit as closely as possible) a given set of probability-quantile pairs. It is constructed entirely from monotone transformations of valid quantile functions, provides analytic monotonicity guarantees, has a provably full-rank design matrix and modality control, and empirically matches or exceeds the accuracy of the Metalog distribution across approximately 3,500 Pearson-system distributions. Finally, Chapter 5 provides rigorous proofs of two previously conjectured monotonicity conditions for the truncated Johnson SL and SB Power-Series Expansions. Bickel (2026) developed a new family of quantile functions (JPSE) based on power-series expansions of the already very flexible Johnson quantile functions (Johnson,1949b,a; Hadlock and Bickel, 2017). Bickel (2026) left as open conjectures several conditions on the truncation orders of the JPSE quantile functions in order for them to be valid quantile functions. We provide rigorous proofs for the truncation order conditions, establishing that odd-order truncations define valid quantile functions for all parameter values and placing the JPSE framework on a firm mathematical footing.
DOI: 10.26153/tsw/64405
Author: Colombe, Connor Lee
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7202272053
