Research
My pages on the Harvard Dataverse, google scholar, and IDEAS/RePEc.
Working Papers
Anticipated labour market discrimination and educational achievement (with Andy Dickerson, Anita Ratcliffe, and Bertha Rohenkohl), 2022, Sheffield Economic Research Paper Series (SERPS) no. 2022017. Revise and resubmit, Journal of Economic Behavior & Organization (special issue on Discrimination and Diversity).
Abstract: Some theories suggest that students who anticipate discrimination in the labour market may invest more in easily observable human capital like education, to signal their productivity to employers and reduce the scope for statistical discrimination. Empirical research on this issue has been hampered, however, by a lack of direct information on anticipated labour market treatment. We use data from a unique longitudinal survey of young people in England to link student expectations of facing discrimination in the labour market to subsequent performance in high-stakes exams. Our findings suggest that the anticipation of labour market discrimination is associated with better exam performance, consistent with the view that students are seeking to counteract potential future penalties.Twitter thread here.Publications
On the power of the Conditional Likelihood Ratio and related tests for weak-instrument robust inference (with Frank Windmeijer), 2023, Journal of Econometrics, 235(1), pp. 82-104 (open access).
Power curves of the Conditional Likelihood Ratio (CLR) and related tests are often simulated in a design where the endogeneity features of the model change with β in a way that 1) makes it less suitable for an analysis of the behaviour of the tests in low to moderate endogeneity settings, or when β and the correlation of the structural and first-stage errors have the same sign; and 2) boosts the power of the CLR test at larger values of |β|. We show that the Likelihood Ratio statistic is identical to the t_{0}(\widehat{β}_L)^{2} statistic as proposed by Mills, Moreira, and Vilela (2014), where \widehat{β}_L is the LIML estimator. In a design where you can directly control the degree of endogeneity, we then show that LIML- and Fuller-based conditional Wald tests and the Fuller-based conditional t_{0}^{2} test are more powerful than the CLR test when the degree of endogeneity is low to moderate, and the conditional Wald tests are the most powerful of these tests when β and the degree of endogeneity have the same sign.Twitter thread here; Mastodon thread here.Code to replicate the figures and table in the paper is available from the Journal of Econometrics website in Appendix D Supplementary data. Previous versions: University of Oxford, Nuffield College, Economics Discussion Paper 2020-W09 (revised July 26, 2021) which replaced the original Nuffield discussion paper (August 2020) that also appeared as Sheffield Economic Research Paper Series no. 2020007.The elusive quest for additionality (with Patrick Carter and Raphael Calel), 2021, World Development, 141, 105393
We use simulations to highlight the biases in how a number of estimation methods evaluate the investment additionality of development finance institutions, and propose a probabilistic approach to evaluating additionality.Supplementary appendix.Replication material.Previous versions: Sheffield Economic Research Paper Series no. 2019022 (December 2019), Center for Global Development Working Paper 495 (September 2018).An early version of the paper was summarised by Paddy in a blog post for CGD.Foreign aid and domestic absorption (with Jonathan Temple), 2017, Journal of International Economics, 108, pp. 431-443 (open access)
What can we learn about aid absorption from a combination of a new 'supply-push' instrument for foreign aid and an instrumental variable estimator that filters out unobserved common factors with possible heterogeneous factor loadings?Replication material.Previous versions: University of Bristol Discussion Paper 15/658 (May 2015), CEPR Discussion Paper Discussion Paper 10211 (Oct 2014), CESifo Working Paper 5029 (Oct 2014), CSAE Working Paper 2014-01 (Jan 2014).The fungibility of health aid reconsidered, 2013, Journal of Development Studies, 49(12), pp. 1746-1754
I examine how the presence of off-budget aid affects estimates of the degree of fungibility of health aid reported in an influential Lancet study.Brief summary on the CSAE blog.Replication material.Working Paper.The full debate: Lu et al. Lancet paper; response by Dieleman, Graves and Hanlon to my JDS paper; my rejoinder.Is foreign aid fungible? Evidence from the education and health sectors, 2013, World Bank Economic Review, 27(2), pp. 320-356
I construct sectoral aid disbursements to estimate the degree of fungibility of education and health aid. I outline how fungibility estimates can be distorted when some aid is off budget, and propose some fixes.Supplementary appendix.Featured on the Roving Bandit blog and the Center for Global Development's Global Health Policy blog.Replication material.Pre-publication version. Working Paper (with GMM results).Measuring and explaining government efficiency in developing countries (with Glenn Rayp), 2007, Journal of Development Studies, 43(2), pp. 360-381
We estimate government expenditure efficiency for a small sample of developing countries and use a general to specific approach to identify its main correlates.