Abstract: Stochastic frontier models have attracted significant interest over the years due to their unique feature of including a distinct inefficiency term alongside the usual error term. To effectively separate these two components, strong distributional
assumptions are often necessary. To overcome this limitation, numerous studies have sought to relax or generalize these models for more robust estimation. In line with these efforts, we introduce a latent group structure that accommodates heterogeneity across entities, addressing not only the stochastic frontiers but also the distribution of the inefficiency term. This framework accounts for the distinctive features of stochastic frontier models, and we propose a practical estimation procedure to implement it. Simulation studies demonstrate the strong performance of our proposed method, which is further illustrated through an application to study the cost efficiency of the U.S. commercial banking sector.
Empirical evidence on the dynamics of investment under uncertainty in the US (2021) with Qazi Haque and Leandro M. Magnusson in Oxford Bulletin of Economics and Statistics
Life-cycle Implications of Latent Health Type Learning
Abstract: Life-cycle models that feature health risk and latent health type implicitly assume that agents are equipped with the knowledge about their type. In this paper, we study the lifecycle decisions and welfare implications of a relaxation to this assumption by introducing type uncertainty for agents, but allow them to resolve it via Bayesian updating of beliefs. We develop a type learning mechanism and document its properties. Then, we embed the type learning mechanism into life-cycle model that features health risk to study the quantitative implications on the life-cycle profiles and welfare of agents facing type uncertainty. When the model is solved numerically using the estimated type-dependent transition probabilities of U.S. high school graduates, our simulation results show that: (i) most agents concentrate their beliefs toward true type within their lifetime, thereby resolving type uncertainty
on average by approximately 80% based on a measure of conditional entropy, and (ii) type uncertainty introduces an additional source of heterogeneity in decisions of agents that ultimately affect life-cycle profiles of agents asymmetrically dependent on their type. Overall, our welfare analyses indicate that the lack of knowledge about their type is costly for agents and can amount to approximately 3% of lifetime consumption.
Estimation of Growth Convergence using a Stochastic Production Frontier Model with Latent Group Structures
Abstract: We test whether countries' growth converges to a common, world production frontier taking into account heterogeneity across countries and over time. We estimate the production frontier and inefficiency term using the recently proposed method of Tomioka, Yang and Zhang (2024). When the production function is characterized by a Cobb-Douglas function, we find evidence in favor of a common, world production frontier and some evidence of growth convergence to that world production frontier. However, under more flexible Translog function, there is evidence of heterogeneity in the frontier and growth convergence result seems to disappear. In both specifications, we find evidence of a mixture structure in the inefficiency term.
Tax Effort and Inequality in Low Income Countries with Dawit Tessema and Abebe Shimeles (IMF internship project)
The Econometric Society, Australasia Meeting, (Melbourne, 2024)
Australian National University, (Canberra, 2021-2024)
The Econometric Society, North American Summer Meeting, (Seattle, 2019)
The Econometric Society, Australasia Meeting, (Auckland, 2018)
Society of Nonlinear Dynamics and Econometrics, (Tokyo, 2018)
Australian Conference of Economists (Sydney, 2017)
University of Western Australia (Perth, 2016, 2017)