Addressing the climate crisis stands as a paramount challenge for sustainable development. Recognizing the gravity of this threat, major nations worldwide have declared 'Net-zero' targets to limit the rise in global average temperature to within 1.5°...
Addressing the climate crisis stands as a paramount challenge for sustainable development. Recognizing the gravity of this threat, major nations worldwide have declared 'Net-zero' targets to limit the rise in global average temperature to within 1.5°C relative to pre-industrial levels. To achieve these ambitious goals, governments are implementing diverse climate policies, including carbon pricing mechanisms—such as emissions trading schemes and carbon taxes—and support for low-carbon R&D. While these interventions are expected to fundamentally reshape energy consumption, industrial production, and supply chains, the academic debate regarding their economic impact remains unsettled. The discourse is currently polarized between the 'Porter Hypothesis,' which posits that environmental regulation stimulates innovation and offsets compliance costs, and the traditional economic view that such regulations impose burdens that erode productivity and competitiveness. Consequently, a rigorous empirical examination of how climate policy influences production efficiency and productivity from a sustainable growth perspective is timely and essential.
Existing studies on environmental production efficiency and productivity have generally incorporated pollutants as undesirable outputs within production frontier analysis. However, they frequently fail to adequately address the endogeneity problem inherent in the selection of production inputs. Endogeneity arises when unobserved production determinants—known to the firm but invisible to the econometrician—influence input decision-making processes. Neglecting this issue risks introducing bias into the estimated production frontier and reducing the precision of policy effect analysis. Therefore, this study aims to construct an analytical model that mitigates input endogeneity and to utilize this framework to analyze the impact of climate policy on environmental production efficiency and productivity.
This research establishes a quasi-Bayesian panel model by combining a control function approach to mitigate endogeneity with Stochastic Frontier Analysis (SFA) based on Bayesian Markov Chain Monte Carlo (MCMC) inference. The methodological distinctiveness and academic contributions of this study are threefold. First, it explicitly addresses the potential endogeneity of input variables—a factor often overlooked in environmental efficiency analysis—by applying the control function approach of Amsler et al. (2016). This enhances the statistical robustness and validity of the production frontier estimation by accounting for the correlation between input variables and the error term. Second, to overcome the limitations of linear assumptions in the Battese and Coelli (1995) model, this study applies the 'scaling property' proposed by Simar et al. (1994) and Wang and Schmidt (2002). This property allows inefficiency determinants to influence the scale rather than the shape of the production frontier, thereby reflecting the structural characteristics of inefficiency more accurately. Third, the study ensures the robustness of environmental productivity estimates by simultaneously deriving the Malmquist index and the Hicks-Moorsteen index, utilizing both output and input distance functions.
Empirical analysis of environmental production efficiency across 124 countries from 2005 to 2019 confirmed the validity of the proposed methodology. The model incorporating endogeneity mitigation yielded a lower Deviance Information Criterion (DIC) value—a measure of model fit—compared to conventional models assuming exogeneity, indicating a significant improvement in goodness-of-fit.
Regarding the determinants of inefficiency, the coefficient for carbon pricing stringency (calculated as carbon pricing coverage × carbon price) showed a statistically significant negative value, implying a reduction in inefficiency. This trend remained consistent regardless of whether endogeneity was controlled, suggesting that stricter carbon pricing—achieved through broader coverage and realistic pricing—acts as a driver for efficiency improvement, increasing economic output while reducing greenhouse gases under identical input conditions. This result empirically supports the strong version of the Porter Hypothesis. However, an examination of carbon pricing data reveals that approximately 45% of current schemes operate at prices below 10 USD/tCO2eq, and price volatility has increased since 2005. Such low prices and high uncertainty may constrain the effectiveness of the price signal mechanism.
Environmental productivity was estimated for 28 OECD countries over the 2005–2019 period. Results indicate that 20 of these nations experienced improvements in both the Malmquist Environmental Productivity Index and the Hicks-Moorsteen Environmental Energy Productivity Index in 2019 compared to 2005. Notably, Greece and Italy demonstrated superior productivity gains, driven by simultaneous improvements in technical change and efficiency change. Conversely, South Korea’s performance was estimated to be relatively poor compared to its OECD peers. While Korea showed slight improvement in the Hicks-Moorsteen index, the magnitude was in the lower tier of the OECD, and the Malmquist index deteriorated due to a slowdown in efficiency change. Of particular concern is the finding that Korea’s technical change index fell below 1.0 during the latter part of the analysis period (2017–2019), signaling a regression in technology. This suggests that Korea's environmental productivity gains have relied heavily on efficiency catch-up rather than technological innovation, highlighting an urgent need to pivot towards innovation-driven strategies.
The analysis of climate policy impacts confirmed the efficacy of a policy mix combining carbon pricing with government R&D investment. The interaction term between carbon pricing and public R&D investment in the environment and energy sectors had a statistically significant positive effect on the Hicks-Moorsteen Environmental Energy Productivity Index. This empirically verifies that the simultaneous operation of regulatory policies (carbon pricing) and support policies (R&D investment) generates synergistic effects in terms of environmental management performance.
Based on these findings, this study concludes that well-designed climate policies can enhance environmental management performance, including production efficiency and productivity. Policy recommendations are as follows: Given that carbon pricing stringency drives efficiency, nations should elevate the qualitative level of their schemes, considering their maturity. Policy coverage should be expanded, and prices rationalized in phases to ensure price signals permeate the entire economy, while accounting for economic conditions and social acceptance. Furthermore, conflicting policy factors, such as energy subsidies that distort price incentives, must be rigorously reviewed.
Nations must also design policies that induce technological innovation. The slowdown in environmental production efficiency improvements in advanced economies (OECD) and the narrowing gap with developing nations suggest that efficiency-driven gains may be reaching a saturation point. Since this study confirmed the validity of the carbon pricing and R&D policy mix, policymakers should actively consider such combinations. However, careful coordination is required to define clear targets and anticipate spillover effects to prevent trade-offs and ensure complementarity.
For South Korea, future climate policy should be designed to incentivize a structural transition toward eco-friendly industries and the qualitative enhancement of green R&D. Since 2018, Korea’s environmental production efficiency has fallen below the global average, and despite high R&D expenditure, the country appears to have entered a phase of technological regression since 2017. These results likely stem from a continued structural reliance on carbon-intensive industries (e.g., steel, petrochemicals, semiconductors) and the insufficient technological yield of government R&D. Future policies must therefore be oriented toward simultaneously promoting industrial restructuring and technological innovation.
While this study attempts to enhance the robustness of analysis through methodological improvements in the stochastic frontier model, several limitations remain to be addressed in future research. First, the assumption of a half-normal distribution for inefficiency prevented the estimation of semi-elasticities for policy effects. Second, the study did not employ the 4-Component Stochastic Frontier Model (4CSFM), precluding a deeper decomposition of inefficiency into persistent and transient components. Third, potential endogeneity arising from the correlation between inefficiency determinants and the error term remains unresolved. Fourth, the R&D data utilized encompasses broad technologies (including water and air quality) and even some fossil fuel-related technologies, limiting the precision of climate policy effect measurement. Finally, the non-application of the directional distance function meant that asymmetric changes between desirable and undesirable outputs were not considered. Future studies incorporating diverse inefficiency distributions (e.g., truncated normal), 4CSFM, endogeneity correction for policy variables, directional distance functions, and refined R&D data will contribute to more precise policy design for realizing a carbon-neutral society.