Researcher Project title Description
Christopher O'Donnell The Estimated Minimum Cost of Operating Australian Electricity Distribution Networks In April 2015, the Australian Energy Regulator (AER) ruled that NSW and ACT electricity distributors were cost inefficient and must take real cuts of between 17% and 31% in their planned revenues. This ruling was informed by econometric estimates of minimum cost. This project uses data provided by the AER to assess the robustness of these econometric estimates. The estimates are found to be highly sensitive to various choices involving both the data and the econometric model. The AER has used estimates of minimum cost that can only be obtained by (a) including an irrelevant variable in the model, (b) omitting a relevant variable from the model, (c) ignoring data on all large firms, and (d) assuming that a time-varying function of timevarying variables does not, in fact, vary over time.
Christopher O'Donnell Measuring and Explaining Productivity Change in U.S. Manufacturing This project uses U.S. manufacturing data to estimate the parameters of a set of industry-specific stochastic production frontier models. The estimated parameters are subsequently used to decompose a set of proper productivity index numbers into estimates of technical change, environmental change, and different types of efficiency change. The results indicate that productivity change in most U.S. manufacturing industries has been driven by technical progress (i.e., the discovery of new techniques for transforming inputs into outputs) and scale-and-mix efficiency change (i.e., changes in economies of scale and substitution). There is no statistical support for many of the assumptions underpinning the growth accounting approach to productivity analysis (e.g., constant returns to scale, no technical inefficiency).
Christopher O'Donnell The Effects of Environmental Change on U.S. Agricultural Productivity: 1960 to 2004 This project uses econometric methods to decompose proper productivity indices into measures of technical change, environmental change, technical efficiency change and scale-and-mix efficiency change. For a finer decomposition, the project assumes that U.S. farmers choose inputs and planned outputs to maximise expected profits in the face of uncertainty about output prices and selected characteristics of the production environment (e.g., rainfall, temperature). The measure of scale-and-mix efficiency change is then decomposed into measures of technical change, expected environmental change, expected output price change and input price change.
Christopher O'Donnell Efficiency Analysis in the Presence of Demand Uncertainty (with Tarmo Puolokainen) This project considers a service industry in which a principal supplies inputs to agents in different jurisdictions in the face of uncertainty about the services that will be demanded in each jurisdiction. The principal aims to minimise the cost of the inputs supplied. The agents aim to use their allocated inputs to provide the services demanded in their jurisdictions. Under-resourcing of agents often leads to unacceptable levels of service. This can reflect badly on both the principal and the agents. We are interested in the cost efficiency of the principal, any under-resourcing of the agents, and the technical and mix efficiency of the agents. In an empirical illustration, we find evidence that 10-25% of rescue service brigades in Estonia are under-resourced. We also find that minimum service levels could have been met in all jurisdictions for 30% (≈ 10 million euros) less than it cost in 2015.
Christopher O'Donnell Maize Productivity and Input Subsidies in Malawi: A State-Contingent Stochastic Production Frontier Approach (with Stein Holden) This project makes cross-sectional comparisons of productivity in a risky agricultural setting. The project develops a new productivity index that satisfies important axioms from index number theory (e.g. transitivity, proportionality). The index can be computed without any information on output or input prices. However, it cannot be computed without an estimate of a statecontingent production frontier. The project uses maximum likelihood methods to estimate a state-contingent stochastic production frontier that explicitly allows for variations in input quality. The empirical results suggest that differences in productivity are due to differences in the production environment and differences in scale-and-mix efficiency. Differences in scale-mix efficiency may be partly driven by variations in access to input subsidies. The maximum likelihood estimator appears to do a poor job of disentangling the effects of technical inefficiency and statistical noise.
Christopher O'Donnell On the Convexity Properties of Group-Specific Cost Frontiers and Cost Metafrontiers (with Kristiaan Kerstens and Ignace van der Woestyne) Metafrontier models are often used in situations where managers in different groups choose outputs from different output sets. In this project, the union of these so-called group-specific output sets is referred to as an output metaset. Under weak regularity conditions, group-specific output sets can be represented by group-specific cost frontiers, and output metasets can be represented by cost metafrontiers. This project discusses the convexity properties of these different sets and frontiers. It then uses parametric and nonparametric estimators to illustrate the consequences of estimating group-specific cost frontiers and cost metafrontiers under convexity and/or non-convexity assumptions that are false.
Christiern Rose Identification of Peer Effects through Social Networks using Variance Restrictions This paper explores the potential for identification of social interactions models based on restricting the variance/covariance of unobserved individual heterogeneity.
Christiern Rose Optimal Injection Points for Information Diffusion This paper considers optimal injection points for diffusion of information about an economic policy through social networks.
Christiern Rose High Dimensional Instrumental Variables Regression and Confidence Sets (with Eric Gautier and Alexandre Tsybakov) This paper studies identification, estimation and inference under sparsity in high dimensional instrumental variables models.
Christiern Rose Identification of Spillover Effects using Panel Data This paper studies identification of spillover effects when the network through which they operate is assumed to be sparse.
Christiern Rose Inference on Social Effects when the Network is Sparse and Unknown (with Eric Gautier) This paper studies estimation and inference on spillover effects when the network through which they operate is assumed to be sparse.
Christiern Rose The Role of Networks in Adoption of Innovation in Healthcare: the Case of Surgery for Colon Cancer (with Carol Propper, Marisa Miraldo and Eliana Barrenho) This paper studies the role of consultant surgeons’ networks in the diffusion of best surgical practice.
Christiern Rose An analysis of the impact of physical activity on cognition among older people in Europe (with Brenda Gannon and Sabrina Lenzen) This paper seeks to identify the the causal effect of physical activity on cognition, accounting for the dynamic nature of cognition and measurement error in self reported physical activity.
Christiern Rose Variable selection in DEA using cardinality constraints (with Antonio Peyrache and Gabriela Sicilia) This paper introduces cardinality constraints for 10 regularized data envelopment analysis, applicable in settings in which the number of inputs and outputs is large relative to the sample size.
Christiern Rose The War on Drugs: An Analysis of the Effects of Supply Disruption on Prices & Purity This paper studies the impact of supply disruption on price and purity of illicit drugs, accounting for their nature as experience goods.
Clement Tisdell Resource and Environmental Economics: Modern Issues and Applications - Second edition The second edition will include a new chapter on the supply of energy and mineral resources as well as new material for chapters on agriculture, agricultural policies and the environment, water management, global warming,energy use and air pollution, recycling and solid waste management, and environmental health economics.
Dong-Hyuk Kim How Costly to Sell a Company: A Structural Analysis of Takeover Auctions To explain why sellers in takeover auctions limit bidders' entry, we structurally measure economic costs incurred by the seller when inviting an additional bidder. In particular, our auction model allows for bidders to discount their synergy values when their rivals obtain the confidential information -- we refer to the induced revenue loss as the information cost. We establish the identification of the model primitives allowing for unobserved heterogeneity, as the confidential information is latent by nature. From a sample with 287 M&A deals of U.S. public companies, we find that the unobserved heterogeneity explains 91.1% of the variation of the synergy value after controlling for the observed target covariates and bidders lower their values by 11.9% for each rival. Our counterfactual analysis reveals that the seller's economic costs are substantial, for example, the information cost and operation cost respectively amount to 1.3% and at least 4.1% for a target with the average market value in the consumer industry.
Dong-Hyuk Kim Empirical Framework for Demand System with High Dimensional Characteristics The project proposes an empirical method to analyse demand systems of markets with highly differentiated products, each characterised by high dimensional covariates. Due to technical difficulties, researchers have been required to focus a few covariates or to impose strong assumptions on consumer preferences, although thorough market data are available for complex products, e.g., panel TVs, smartphones, etc. Our method will be designed for such big data with minimal restrictions and thereby making it possible to flexibly analyse demand structure with full information. Our technical innovation and applications would result in a series of high-quality publications and help industry players, such as firms and government, make better decisions.
Dong-Hyuk Kim Bidder Asymmetry and Heterogeneous Risk Aversion This project studies identification and estimation of first price auctions where the marginal valuation distribution as well as utility function are allowed to be bidder specific. We show that the model primitives are identified when all submitted bids are observed and the bidder configuration exogenously varies. Our identification result extends to the model with a certain form of auction specific unobserved heterogeneity. We develop an empirical method to estimate the model primitives and show its performance using a series of Monte Carlo experiments.
Knox Lovell Indirect Productivity Measurement There is a long but largely neglected tradition of using prices rather than quantities to estimate productivity change. We revisit this tradition, and we propose a novel analytical framework for indirect (aka dual) productivity measurement.
Knox Lovell Circularity It is well known that most popular empirical price, quantity and productivity indexes, and the popular Malmquist quantity and productivity indexes, are not globally circular. We show that each of these indexes satisfies the circularity property locally.