This paper develops a multi-ware data envelopment analysis framework for measuring land requirements in livestock systems where production and pollution abatement are jointly feasible but technologically distinct. The model imposes free disposability only on land and locally matches all other netputs. This avoids imposing monotonicity and allows each netput to be land-using or land-saving at the observed farm state. The feasible technology is represented as the upper envelope of production and proactive abatement land requirements. Shadow prices are then recovered from the tighter land constraint. This structure reduces the attenuation and sign inconsistencies found in pooled or unconditional frontier models. We apply the model to regulated commercial hog farms in China using aerobic-anaerobic treatment systems from 2012 to 2014. About 64 percent of farms are abatement-binding. Abated NH3-N is consistently land-using, while several abatement inputs are locally land-saving. This suggests that treatment capacity can reduce the marginal land burden of compliance. We also introduce two local policy diagnostics. Inefficiency measures excess land use relative to the joint upper envelope. Imbalance measures the gap between production and abatement land requirements at the same observed netput bundle. The results reveal substantial heterogeneity across regimes and regions, with important implications for environmental policy design.
Production costs are often evaluated beforehand by engineers, given the main characteristics of the output and using the decomposition of the production process in elementary tasks. The economist usually observes ex post the cost and the characteristics of the product. In this paper we propose to take advantage of the situation where we have information from both approaches. We consider a system of two equations, one being a standard regression model (for the technical cost function) and one being a stochastic frontier model for the economic cost function where a positive part is explicitly introduced to model the difference between the two costs. We derive estimators of this joint model and derive its asymptotic properties. The models are presented in classical parametric approach, with few assumptions on the stochastic properties of the joint error terms. We also suggest a way to extend the model to a nonparametric approach, the latter provides an original way to model and estimate nonparametric stochastic frontier models. The techniques are illustrated in the case of the cost function for the distribution of gas in France.
The emergence of the new public management paradigm, the growing number of students, and intensified international competition for external funding in large research projects have pressed universities to expand both the size and role of non-academic staff (i.e., administrative and technical personnel) to support the academic staff in addressing these increasingly complex challenges effectively. While the contribution of academic staff to university performance is well established, professors engage in teaching, research, and knowledge transfer activities, the role of non-academic staff remains more ambiguous. Their impact can be beneficial if they facilitate and enhance the core activities of academic staff. However, if their presence leads to greater administrative burdens, diverting time away from teaching and research, their effect may be detrimental. We provide new insights into this critical issue by examining European universities. We account for the heterogeneity of non-academic staff as an “unobserved" factor and estimate it using a nonparametric approach. We then evaluate its impact on university performance across Europe. Our empirical analysis, based on a sample of 401 European universities, reveals that the proportion of non-academic staff does not exert a statistically significant effect on university efficiency. In contrast, their unobserved heterogeneity, capturing qualitative and structural differences beyond mere size and composition, emerges as an important determinant of university performance. The latent factor significantly influences both the efficient frontier and the distribution of inefficiencies. Universities exhibiting very high levels of staff heterogeneity combined with large administrative shares tend to experience reduced efficiency, consistent with patterns of bureaucratic overload. Conversely, universities that effectively manage and integrate diverse professional competences within their non-academic workforce are able to transform heterogeneity into a productive resource, enhancing both teaching and research efficiency across Europe.
This paper adopts KLEMS method to measure TFP and decomposes the economy-wide resource reallocations into the growth rate of factor inputs and service prices at industry level to explore the contribution of individual industries to the overall resource reallocation and its causes, and takes the United States and Japan as references to explore the lessons for the development of China’s TFP. The combined reallocation of resources dominated by significant positive labor reallocation contributes 36
As artificial intelligence (AI) spreads worldwide, its impact on labor markets is still unfolding and remains uncertain, yet potentially far-reaching in developing economies undergoing rapid structural transformation. This paper provides the first large-scale evidence from China, where AI investment and adoption have expanded rapidly, linking local AI labor demand to individual wage outcomes. We construct city–year measures of AI labor demand from 1.6 million online job postings between 2016 and 2024, capturing the intensity, breadth, and diversity of AI-related hiring, and merge them with nationally representative microdata from the China Family Panel Studies (2016–2022). Fixed-effects estimates show that local AI labor demand has positive impacts on individual wages: a one-unit increase in AI demand (1,000 postings, firms, or job titles) raises wages by about 0.2–0.3
Improving farm efficiency is central to enhancing productivity, income, and structural transformation in European agriculture. However, inefficiency is often persistent and rooted in structural constraints. This paper examines how the Common Agricultural Policy (CAP) affects technical efficiency and its persistence in animal farming, using Slovenia as a case study. Slovenia provides a relevant empirical setting due to the dominance of small-scale farms, strong reliance on subsidies, and limited economies of scale – characteristics shared by many structurally constrained EU Member States. We apply a Bayesian dynamic stochastic frontier model that jointly accounts for efficiency persistence and technological heterogeneity. Using disaggregated CAP subsidy data, we assess how different policy instruments influence short-run and long-run technical efficiency across heterogeneous farm types. The results reveal four main findings. First, there is strong evidence against a common production frontier, indicating substantial technological heterogeneity among animal farms. Second, the effect of CAP subsidies on short-run technical efficiency depends both on the direction of their impact on efficiency persistence and on the level of technical efficiency in the previous period. Similarly, their impact on long-run technical efficiency is shaped by their effect on persistence and by the level of long-run efficiency. Third, the effects of CAP subsidies are heterogeneous across instruments: investment and other subsidies enhance both short-run and long-run technical efficiency, whereas decoupled payments and agri-environmental subsidies reduce efficiency at both horizons. Fourth, payments for less-favored areas consistently lead to a deterioration in technical efficiency in both the short and the long run.
Flexible work arrangements are increasingly becoming the norm across many countries, particularly in developed nations such as Canada and the United States. Remote work (work-from-home, or WFH, for short) has gained significant popularity in recent years, especially after the outbreak of the COVID-19 pandemic in March 2020. Building on seminal contributions by Bloom et al. (2015) and Davis et al. (2024), we develop a simple theoretical model that captures the trade-off between externalities associated with remote work and those derived from in-office work. The model is based on the premise that both productivity and firm profitability are influenced by the intensity or proportion of remote work arrangements by firms. Ultimately, higher workforce productivity and firm profitability due to remote work translate into higher wages. By considering different collaboration structures among employees, the model demonstrates that while remote work can lead to cost savings and positive externalities, excessive adoption may undermine benefits, increase management complexity, and raise the risk of employee shirking. The central theoretical result is an inverted U-shaped relationship between the extent of remote work and wages. To test this prediction, we use data from the Canadian Labour Force Survey to examine the relationship between industry-level remote work intensity and individual wages. The empirical findings reveal that wages rise with remote work intensity up to a threshold, approximately 52.1–63.9
This paper develops a new primal approach for the estimation of marginal costs under conditions in which technical efficiency is explicitly incorporated into the cost evaluation. To do this, we exploit the dual relationship between the input distance function and the cost function. We apply this framework in order to analyse the impact of hospital waiting lists on efficiency and costs. Waiting lists have a dual effect on social welfare. On the one hand, delays in healthcare provision exacerbate patients’ health problems. On the other hand, redistributing patient care from periods of high demand to those of lower demand allows the healthcare system to operate closer to full capacity, thereby enhancing hospital technical efficiency. However, when waiting lists become excessively long, this effect diminishes, as idle times are eliminated. Moreover, patients’ health may deteriorate significantly, leading to increased treatment costs that may outweigh the savings associated with reduced idle time. In this context, we estimate the marginal cost of hospital output while accounting for waiting lists and their impact on technical efficiency. Our results indicate that, while moderate waiting lists can improve productive efficiency and reduce costs, excessively long waiting lists may negate these benefits.
High technology products are characterized by the rapid introduction of new models and the corresponding disappearance of older models. The paper addresses the quality adjustment problem associated with the construction of price indexes for these products. A main method for dealing with this problem is the use of hedonic regression models. Hedonic regressions use either product characteristics as explanatory variables (Time Dummy Characteristics regressions) or the product itself as the ultimate characteristic (Time Product Dummy regressions). The paper considers weighted and unweighted Time Product Dummy regressions. The indexes which were generated by the hedonic regressions are compared to traditional index numbers that did not make any special adjustments for quality change. The Expanding Window variant of a Weighted Time Product Dummy regression was used to address the chain drift problem and the problems associated with extending a series that cannot be revised. Finally, the estimation of systems of inverse demand functions was also used to generate various price indexes. Eighteen alternative approaches were implemented using Japanese price and quantity data on laptop sales for the 24 months over the years 2021–2022.
Existing research on the impact of economic agglomeration on carbon emission efficiency has yielded inconsistent results. This study is among the first to use a spatial stochastic frontier analysis model that addresses both spatial effects and endogeneity to examine the impact of economic agglomeration on carbon emission efficiency using a panel dataset of Chinese cities. The findings indicate that an increase in economic agglomeration facilitates the reduction of carbon emissions and enhances carbon emission efficiency, aligning with sustainable development policies. Further counterfactual analysis suggests that targeted agglomeration policies yield diminishing marginal environmental returns. Specifically, increasing economic agglomeration in less agglomerated, low-density cities unlocks massive carbon emission reduction potential strictly through efficiency enhancements. Based on this analysis, it is recommended that policymakers in China avoid one-size-fits-all approaches and instead take targeted measures, such as eliminating rigid administrative barriers like the hukou registration system to build unified large markets, to prioritize population concentration in less agglomerated, low-density areas, thereby unlocking disproportionately large carbon emission efficiency dividends.
The papers in this Collection were presented in the Special Plenary Session in Honor of Professor Finn R. Førsund, June 20, 2024, at the XVIII European Workshop on Efficiency and Productivity Analysis (EWEPA). During the plenary session, he received the Lifetime Achievement Award for his significant contributions to the science of productivity and efficiency analysis, recognized by scholars, practitioners, and students from the efficiency and productivity community. The award was presented to Professor Førsund by EWEPA and the International Society for Efficiency and Productivity Analysis (ISEaPA).
Mexico’s Production for Wellbeing program (Programa de Producción para el Bienestar, PWp) aims to increase the production and productivity of grains, coffee and sugarcane to enhance domestic food self-sufficiency and crop competitiveness for the nation’s small and medium-sized farmers. Through stochastic production frontier analysis (Kumbhakar and Lovell 2000) and meta-frontier (Battese and Rao 2002; Huang et al. 2014) analyses, this study investigated whether the PWp has influenced the technical efficiency (TE) of those beneficiaries who produce white corn or sugarcane, considering an impact evaluation framework, using the Mahalanobis distance for matching (Rubin 1980). Microdata come from the National Agricultural Survey (INEGI, 2019), and are complemented with poverty, climate, and regional data to identify microregions for policy intervention (Maruyama et al. 2018). We find that PWp beneficiaries were approximately 17.67
Evidence from farm trials and farm household surveys indicates that adopting Bacillus thuringiensis (Bt) cotton can significantly increase cotton farm productivity in developing countries. Whether these gains scale up to the industry level, however, remains unclear. This paper employs the regression control method (RCM) within a cross-country comparative framework to estimate the impact of Bt cotton adoption on aggregate agricultural productivity in China. Drawing on a balanced panel of 86 non-Bt-cotton-producing countries and regions, we construct counterfactuals for China and evaluate the effects on cotton yield and agricultural total factor productivity (TFP). Our results show that Bt cotton adoption exerted positive effects on industry-level cotton yield, but its effects on agricultural TFP were both statistically insignificant and economically small. Further analyses suggest that limited resource reallocation toward larger, more efficient cotton farms after Bt cotton adoption dampened aggregate gains, highlighting structural constraints that impede the translation of micro-level gains from biotechnology into sector-wide productivity improvements.
In this work I briefly describe my views on the importance, challenges and solutions for the estimation and inference of the aggregate productivity and efficiency scores, and related indexes and indicators, designed to represent various economic systems: firms, hospitals, industries, countries and their groups or distinct sub-groups. This article also serves as an indirect tribute for Professor Léopold Simar, as I also point out how he has substantially shaped and navigated this stream of literature.
Theory suggests that relatively inefficient firms should have lower and more uncertain future cash flows than their efficient peers, which should lead to lower current equity values and higher future returns. However, we find that inefficient firms experience significantly lower returns than their more efficient counterparts. We provide evidence that a possible reason for the negative drift in returns is that investors are not fully aware of firms’ operational inefficiencies and are negatively surprised when future negative earnings are announced. Furthermore, we document that analysts do not properly incorporate information about operational inefficiency into their earnings forecasts and target prices and seem to overlook the issue of operational efficiency during conference calls, particularly for inefficient firms.
The stochastic frontier model (SFM) is widely employed in the analysis of productivity and efficiency, yet strict parametric forms, such as the Cobb-Douglas and Translog functions, are often assumed for modeling production, leading to potential misspecification issues. While semi- and nonparametric SFMs offer greater flexibility, they face challenges in imposing monotonicity and concavity to maintain their desirable economic interpretation. We develop a framework which enforces the shape restrictions within deep neural networks (DNNs). The stochastic frontier model we develop (DNN-SFM) leverages the flexibility and predictive power of DNNs while preserving key properties of a production function, such as free disposability and diminishing marginal product. Additionally, we demonstrate how to use Shapley values to measure and interpret global and local effects of individual inputs on the production frontier in cases when model parameters to not admit a simple interpretation. The performance of the proposed method is assessed using simulations while a real-world application to rice production in the Philippines illustrates empirical relevance of the proposed method.
The Benefit of the Doubt (BoD) models offer a flexible way to construct composite indicators by endogenously aggregating performance metrics, appealing to managers who favour indices over complex DEA frameworks. While prior studies focus on BoD’s aggregation mechanism and employ weight restrictions to avoid zero weights, few leverage peer benchmarks and target values to enhance performance insights. This paper advances the BoD approach in three ways. First, we provide a comprehensive review of input- and output-oriented BoD model variations, clarifying which formulation best suits scenarios with (1) multiple outputs and a single input normalised to one, or (2) multiple inputs and single output set to one, thereby ensuring meaningful and actionable benchmarks. Second, we develop models that identify the minimum weight bounds (α) each decision-making-unit (DMU) can sustain while preserving its unconstrained efficiency score, with higher α indicating greater robustness across all performance dimensions. Third, we propose using the lowest α among top-performing (strong efficient) units as a universal weight bound, maintaining the efficiency frontier while exposing underperformance (non-radial slacks) in weaker units. These innovations strengthen the analytical rigour and policy relevance of BoD, enabling robust evaluation and targeted improvement strategies. We illustrate the approach using data from 30 large U.S. banks, evaluating performance across four prudential pillars—capital adequacy (Tier 1), earnings (ROE), operating efficiency, and funding stability—to derive interpretable scores, peer benchmarks, and actionable priorities.
This paper proposes a novel nonparametric panel data framework for estimating conditional production frontiers and efficiency measures that explicitly accounts for spatial interdependencies. By integrating recent advances in nonparametric frontier estimation with spatial panel data analysis, the proposed approach offers a flexible and robust framework for assessing productivity and efficiency in the presence of spatial interactions, explicitly accounting for both global and local spatial effects. By extending recently developed tools for estimating Malmquist productivity indices to conditional nonparametric frontier efficiency models, we provide a refined decomposition of productivity growth into technological change, efficiency change, and scale effects within a fully nonparametric framework. Applying this framework to a comprehensive dataset on European regions, we provide new evidence on spatial patterns of productivity growth and efficiency dynamics across the EU. The results reveal marked heterogeneity in regional performance and highlight the crucial role of spatial spillovers in shaping productivity outcomes. Ignoring these interdependencies can lead to mismeasurement of productivity trends, reinforcing the value of our proposed spatial nonparametric frontier approach for policy and performance analysis.
