Imperial Research Proposal Example: Energy policy experiments researcher to market reform (Score 93)

Electricity market reform in liberalised systems rarely proceeds through single legislative acts. It advances through sequences of bounded interventions — capacity mechanism adjustments, retail price deregulation pilots, demand-response trials — each generating observable market responses before the next design iteration. Despite this, the academic literature on energy policy evaluation tends to treat reforms as discrete before-and-after events rather than as iterative experiments embedded in ongoing market structures. This proposal asks: under what institutional and market conditions do policy experiments in electricity markets produce durable changes in wholesale price formation and supplier behaviour, and which experimental design features predict whether those changes persist after the intervention ends? The question is bounded in two ways. First, it restricts attention to electricity markets in OECD jurisdictions that have undergone partial or full liberalisation, where both the intervention and the market outcome are observable through publicly reported data. Second, it focuses on the mechanism linking experimental design features — duration, scope, reversibility, and participant self-selection — to downstream market outcomes, rather than evaluating any single country's reform trajectory. This framing yields transferable analytical findings while remaining feasible within a one-year MSc research project. The practical motivation arose from work on a regulatory memo examining how a demand-response experiment interacted with existing balancing market rules, producing price signals ambiguous to both generators and the system operator. That experience made the gap between experimental intent and market-level outcome concrete, raising a question the memo could not answer: whether the ambiguity was design-specific or reflected a more general structural feature of how policy experiments interact with incumbent market arrangements. Two bodies of scholarship bear on this question and have developed largely in parallel. The first is the economics of electricity market design, associated with work on nodal pricing, capacity adequacy, and competitive wholesale markets. This literature — drawing on contributions from Joskow, Stoft, Cramton, among others — is technically precise about equilibrium conditions but tends to evaluate market structures comparatively rather than tracing how incremental regulatory interventions disturb and re-equilibrate market behaviour over time. Reform is modelled as a structural shift, not as a sequence of experiments with feedback loops. The second is the growing literature on policy experimentation and regulatory learning, drawing on public administration scholarship and quasi-experimental evaluation methods applied to energy programmes. Work on smart meter rollouts, time-of-use tariff pilots, and renewable obligation schemes has become methodologically rigorous, yet it evaluates individual programmes in isolation and rarely asks whether the experimental design itself, independent of policy content, shapes the market response. The gap is specific: no systematic comparative analysis examines how the structural features of electricity policy experiments — duration, reversibility, participant selection mechanism, and interaction with existing market rules — predict whether resulting price or behavioural changes persist once the experiment concludes. Existing reviews either catalogue reform outcomes without isolating experimental design as a variable, or evaluate individual pilots without connecting findings to wholesale market dynamics. This proposal addresses that gap directly. The study combines structured comparative case analysis across a purposively selected set of electricity market experiments from OECD jurisdictions with a quantitative component examining wholesale price series around intervention windows. Phase one: case selection and documentary analysis. Using publicly available regulatory filings, energy authority reports, and academic case literature, I will identify approximately twelve to sixteen policy experiments in liberalised electricity markets between 2005 and 2023. Selection criteria require a defined start and end date; available wholesale price or supplier-behaviour data for pre-, during-, and post-intervention periods; and a published formal evaluation or consultation response from the regulatory authority. This produces a structured dataset of experimental design features coded against a common framework — duration, reversibility, participant selection mechanism, and interaction with balancing or capacity market rules. Phase two: quantitative price-series analysis. For each case where hourly or daily wholesale price data is available through public repositories — including the ENTSO-E Transparency Platform and national regulatory databases — I will apply interrupted time-series analysis to estimate the intervention effect and the rate of reversion or persistence after the experiment ends. Where multiple cases share comparable market structures, a difference-in-differences specification will control for concurrent market-wide shocks. The sample of twelve to sixteen cases is modest; I will be explicit about the limits of cross-case inference, as the goal is pattern identification and hypothesis generation rather than causal identification in the econometric sense. Phase three synthesises qualitative and quantitative findings to produce a typology of experimental design features associated with durable versus transient market effects. This typology — the primary analytical output — will be framed as conditional propositions testable in future work with larger samples, rather than as definitive causal claims. All data sources are publicly available. ENTSO-E, Ofgem, the Federal Energy Regulatory Commission, and equivalent national bodies publish wholesale price data, consultation documents, and programme evaluations as open-access resources. No proprietary datasets, firm-level data, or personal data are required, so the project does not raise significant ethical review concerns beyond standard practice for secondary data analysis. I will confirm the applicable ethics pathway with the department at the outset. The principal feasibility risk is case availability: not all experiments will have published evaluations of sufficient detail to code design features reliably. I will address this by building a reserve list during preliminary scoping, so the final sample of twelve to sixteen can be drawn from an initial list of approximately twenty-five. A secondary risk is data comparability across market structures; I will document structural differences explicitly and treat them as moderating variables rather than assuming equivalence. Provisional timeline: months one and two, literature review and case selection framework; months three and four, documentary coding and dataset construction; months five through seven, quantitative analysis; months eight and nine, synthesis and typology development; months ten through twelve, writing and revision. This schedule is consistent with a one-year MSc project and includes a two-week contingency buffer before submission. Imperial's Centre for Energy Policy and Technology and the research activity associated with the MSc Energy Policy programme provide the methodological and domain environment this project requires. The programme's emphasis on quantitative policy analysis and engagement with regulatory institutions means the combination of structured comparative analysis and time-series methods used here is consistent with the department's research culture. The project requires access to the ENTSO-E Transparency Platform and equivalent national databases, all publicly accessible. It does not require laboratory facilities, proprietary software beyond standard econometric packages, or fieldwork travel. The analytical methods — interrupted time-series and difference-in-differences — are covered in standard postgraduate econometrics training, supplemented through the programme's quantitative methods provision. The expected contribution is modest and deliberately so: a structured comparative typology of experimental design features and their association with market persistence, presented as a resource for regulatory practitioners and researchers working on electricity market evaluation. It does not claim to resolve the causal identification problems inherent in observational market data, but it provides a more systematic framework than currently exists for understanding why some policy experiments leave durable market traces and others do not.