Environmental and Technical Efficiency in Large Gold Mines in Developing Countries

This study evaluates the environmental performance of large gold-mining operations by applying a by-production model that specifies emission-generating technology, while incorporating a four-way error approach that captures mine-size heterogeneity, transient and persistent technical efficiency, and random errors. We applied a true random effects model (TREM), and a simulated maximum likelihood estimator (SMLE) based on the generalised true random effects model (GTREM). The former approach was estimated as a benchmark, while the latter was employed to estimate a four-component panel data stochastic frontier model. The four-components estimate separates firm heterogeneity from persistent and time-varying inefficiencies, thus providing more robust efficiency estimates and policy insights. Firm-level data from 2009 to 2018 were used; the results show the presence of environmental and technical inefficiencies. Moreover, each inefficiency was decomposed into transient and persistent inefficiencies. The GTREM predicts the average inefficiency to amount to 34% environmental (interaction between 19% transient and 18% persistent) and 30% technical (interaction between 4.4% transient and 27% persistent). The transient component of technical efficiency does not change over time, which may imply that the mines’ managerial approaches are static. The presence of technical inefficiency implies that more than the minimal amounts of inputs are used to produce a given level of desirable output, which could be due to moral hazards and asymmetric information such as principal-agent problems. The presence of (environmental) inefficiency in the by-production model means that more than a minimal amount of the undesirable output is produced. The overall environmental performance of the mines in developing countries is low (66%) compared to other sectors, which indicates that there could be structural rigidities, poor environmental policies and regulations, poor enforcement, or any combination of the three. We also found robust empirical ...