Lead research economist — technology, productivity & labour markets, applied econometrics and policy.
I am a PhD economist with more than fifteen years of experience across government, think tanks and academia, working on policy-relevant empirical economics: productivity, workforce transformation, technology adoption and large-scale administrative data.
I currently lead the economics function at the Health Foundation, covering NHS demand, funding, workforce, productivity and technology, and oversee the REAL Supply Research Unit, a partnership with the University of York. My work combines applied and structural econometrics, causal inference, machine learning, macro modelling and policy translation. I am particularly interested in how frontier AI reshapes tasks, occupations, productivity, inequality and public-policy choices.
I have worked across government, academia and think tanks on topics relating to technology and innovation, productivity, contracting in health & care, and applied microeconometrics more generally. I hold a PhD in Economics from Birkbeck, University of London.
Medical innovations and improvements in healthcare have contributed to large improvements in welfare but also growing healthcare costs. At the same time, financial constraints in public sector institutions may affect the delivery of services. We examine whether hospital trusts with budget deficits prescribe new, cost-effective but expensive treatments for Hepatitis C differently. We compile a novel panel dataset of hospital trusts providing acute care in England, linking detailed data on financial statements, workforce statistics, prescribing volumes, hospital activity, and quality of care. We employ two complementary identification strategies: two-way fixed effects and an instrumental variables approach, using historical deficits and hospital activity in large disease groups as instruments for current financial positions. Our findings indicate that hospitals in worse financial positions prescribe fewer treatments: a 10% increase of a standard deviation in a trust's surplus (£2 million) is associated with a 1.7–2.7% increase in prescribing new Hepatitis C medicines. We rule out several potential mechanisms, including staff composition, drug costs, and quality of care.
This paper examines GP prescribing behavior for on-patent diabetes therapies in NHS England and quantifies their short-run impact on secondary-care use. Leveraging a rich practice-level panel and a structural demand framework that treats GPs as perfect patient agents, we instrument net price with exogenous rebate shocks under four alternative specifications. Implied own-price elasticities of uptake lie between −0.2 and −0.6, with smaller practices exhibiting greater sensitivity. Cross-price elasticities between the new-treatment class and legacy therapies are close to zero. A 2SLS IV estimation then links a one-percentage-point increase in the new-treatment share to a 4 percent rise in diabetes-related hospital admissions over two years—equating to roughly 0.25 additional episodes per practice. These results challenge the prevailing view that innovative therapies relieve secondary-care pressures, instead revealing meaningful short-term cost pressures. By integrating structural demand estimation with real-world outcome monitoring, our study offers a practical blueprint for evaluating the system-wide cost-effectiveness of novel treatments in universal health-care settings.
This paper examines how pharmaceutical firms strategically time the launch of new products in response to patent protection and competitive pressure. I estimate a dynamic oligopoly model to assess whether revenue replacement or pre-emption motives primarily drive these innovation decisions. Removing the replacement channel raises launch probability to 79.0%, indicating that replacement motives dominate strategic delay. Conversely, eliminating pre-emption motives increases launch probabilities to 49.9%, suggesting that competitive entry pressures play a secondary but notable role. Overall, extending patent length by five years increases launch probabilities only modestly, from 1.80% to 2.05%, while materially strengthening replacement-driven delay incentives. These findings suggest that policies aimed at fostering pharmaceutical innovation should prioritise reducing replacement incentives rather than extending patent terms.
Interactive tools and visualisations built from public data. Click any card to open the live site.
Data Darbar — Pakistan in Numbers
An interactive district-level map of Pakistan visualising census, PSLM, labour-force and HIES indicators, built on a Python ETL pipeline.
Live site →·Code
UK Election Projection Tool
A constituency-level seat projection model for UK general elections — uniform and proportional swing, tactical-voting scenarios, poll smoothing (Kalman/LOWESS) and an interactive 650-seat map, with a Bayesian hierarchical model underneath.
Live site →·Code
Demopolis London
An interactive map of London election results, exploring results and demographic patterns across the capital.
Live site →·Code
Population Projection Visualiser
A Streamlit app for exploring population projections interactively — scenarios, age structure and demographic change over time.
Live site →·Code
A full curriculum vitae, including experience, education, methods and engagements, is available as a PDF.