published Papers

Staying at Home: Mobility Effects of COVID-19, with John Stromme and Anson Zhou.

• Covid Economics: Vetted and Real-Time Papers, Issue 4, 86-102, April 2020

• Vox EU briefing here.

• Longer version.

Offline training for improving online performance of a genetic algorithm based optimization model for hourly multi-reservoir operation, with Duan Chen, Arturo S. Leon, Claudio Fuentes, and Qiuwen Chen.

• Environmental modelling & software, Volume 96, 46-57, October 2017

Working Papers

• Alternative Asymptotics and Fixed-Alternative Power Analysis (R&R at Econometric Theory)

  • Abstract: When constructing Wald tests, consistency is the key property required for the variance estimator.  This property ensures asymptotic validity of Wald tests and confidence intervals. Classical efficiency comparisons of hypothesis tests indicate all consistent variance estimators lead to equivalent Wald tests under local power approximations.  This paper develops a simple asymptotic framework under fixed alternatives, which leads to new conclusions.  In particular, we identify that variance estimation will have a first-order impact on the efficiency of Wald tests when size tends to zero with sample size while effect sizes are fixed.  We apply this framework to several applications, including cluster-robust inference and quantile regression.  In the case of cluster-robust inference, we provide for an asymptotic framework in which choosing the wrong cluster size can lead to lower power of tests.  Simulations demonstrate that the results are applicable to moderate sample sizes and that the results can provide a useful approximation for size in standard ranges.

• Robust Tests for Heavy-Tailed Data

  • Abstract: There has been increased interest in economics and finance on the effects of heavy tailed data on the behaviour of test statistics and estimators.  First-order large sample theory can be unreliable and lead to inefficient inference when the statistics in question have heavy tails.  This paper makes two contributions.  First, a lower bound for the rate at which Type-II error converges to zero is provided. While classical results on similar rates of convergence of power focus on situations where exponential rates are feasible, we find the rate to be a much slower polynomial rate when the data have heavy tails. Second, a new procedure based on robust generalized empirical likelihood is developed which leads to test statistics with thinner tails.  These tests share the desirable first-order properties that conventional tests, have, while improving their global efficiency properties.  Simulation evidence shows that the efficiency gains from the new testing procedure can be substantial.  An application demonstrates the ability of the testing procedure to detect assets which tend to carry more or less risk than the market in aggregate.

•  Asymmetric Fertility Elasticities, with Anson Zhou and Chong Pang.

  • Abstract: Over the last five decades, a remarkable reversal has taken place where many countries around the world shifted their policy stances from suppressing to maintaining or promot- ing childbirth. Exploiting rich historical data, this paper documents that the effectiveness of pro-fertility policies is much smaller than the anti-fertility ones – a new fact that cannot be explained by existing models with smooth aggregate fertility demand. We then develop a dynamic model where the government minimizes costs due to policy expenditures and fer- tility levels that are either too high or too low. We show that asymmetric fertility elasticities lead to two novel policy implications: First, the cost-minimizing fertility level is higher than the long-run target; Second, fertility levels possess positional values that should be taken into account in policy evaluations. Lastly, we propose a new theory of fertility choice fea- turing loss aversion to provide a micro-foundation of asymmetric elasticities and discuss competing alternatives.

Teaching

University of Exeter: 

• Fall 2022: Econometrics (BEE2031) (with Climent Quintana-Domeque)

• Fall 2023: Econometrics (BEE2031)

Past Teaching: see CV