EXTENDED ABSTRACT
Introduction
The rapid expansion of the digital economy and e-commerce has fundamentally transformed global production, distribution, and consumption paradigms, offering unprecedented opportunities for economic growth and operational efficiency. As physical processes are increasingly replaced by digital alternatives—a phenomenon known as dematerialization—there is a theoretical expectation of reduced environmental pressure. However, the concurrent rise in e-commerce has introduced a complex environmental paradox. While digitalization optimizes supply chain management and reduces inventory waste, it simultaneously exacerbates new environmental challenges, most notably the exponential increase in individual packaging, fragmentation of the last-mile delivery, and elevated rates of product returns. Despite these conflicting outcomes, empirical research quantifying the net effect of business-level e-commerce adoption on internal waste management indicators remains scarce. Most existing literature focuses heavily on macroscopic carbon footprints or qualitative consumer-side packaging issues rather than firm-level operational outcomes. To address this critical gap, the present study aims to quantitatively investigate the impact of e-commerce adoption on waste generation and recycling rates among businesses in 12 selected European Union (EU) member states over the period from 2015 to 2020. Furthermore, the study explores the moderating role of environmental policy stringency (EPS) in shaping these relationships, providing nuanced insights into the interplay between technological adoption and institutional frameworks.
Materials and methods
This research adopts an applied, descriptive-analytical approach utilizing balanced panel data from 12 EU member countries (Austria, Belgium, Czech Republic, Germany, Hungary, Italy, Netherlands, Poland, Portugal, Slovakia, Slovenia, and Spain) spanning from 2015 to 2020. Data were meticulously compiled from highly authoritative international databases, including Eurostat, the World Bank, and the Organization for Economic Cooperation and Development (OECD). The dependent variables are per capita waste generation (WG) and the recycling rate (RR). The primary independent variable is the rate of e-commerce adoption among businesses (ECOM), measured as the percentage of enterprises engaged in electronic sales. The moderating variable is the OECD’s Environmental Policy Stringency (EPS) index. Control variables encompass energy consumption (NRG), gross domestic product per capita (GDP), and population density (POPDENS). Given that official waste management statistics are systematically reported on a biennial basis (even years), a scientifically rigorous linear interpolation method was employed to estimate values for the intervening odd years (2015, 2017, 2019), ensuring a continuous time series while preserving local macroeconomic trends. For the econometric analysis, preliminary diagnostic tests (Chow, Breusch-Pagan, and Hausman) were conducted, which collectively indicated that the Random Effects model was the most appropriate structural specification. However, subsequent tests confirmed the presence of cross-sectional heteroscedasticity and serial autocorrelation within the panel data. To robustly correct for these classical assumption violations, the models were estimated using the Estimated Generalized Least Squares (EGLS) method, fortified with cluster-robust standard errors to ensure the validity of statistical inferences.
Results and discussion
The econometric estimations yielded highly significant and nuanced findings. In the first model (Waste Generation), the direct coefficient of e-commerce adoption was -4.96 (p < 0.01), indicating a strong waste-reducing potential inherent to digital commerce operations. However, the interaction term between e-commerce and environmental policy stringency (ECOM × EPS) exhibited a significant positive coefficient of +4.70 (p < 0.01). Marginal effect analysis reveals a conditional relationship: e-commerce significantly reduces per capita waste generation only in regulatory environments where the EPS index is relatively weak (below a threshold of approximately 2.87). In countries with stricter environmental regulations—which encompasses the sample mean (EPS ≈ 2.99)—the net effect of e-commerce neutralizes and even shifts to a slightly positive value. This dynamic can be explained through institutional isomorphism; under weak regulations, the intrinsic efficiencies of e-commerce (e.g., dematerialization, inventory optimization) create a substantial comparative advantage over traditional retail. Conversely, stringent regulations enforce high resource-efficiency standards across all market participants, eroding the relative advantage of digital platforms while exposing the rebound effects of increased packaging waste associated with granular e-commerce logistics.
A parallel conditional dynamic emerged in the second model (Recycling Rate). The direct effect of e-commerce on recycling was positive (+2.95, p < 0.01), highlighting the capacity of e-commerce reverse logistics to facilitate the collection of recyclable materials. Yet, the interaction coefficient (ECOM × EPS) was highly negative (-2.88, p < 0.01). Consequently, the positive marginal impact of e-commerce on recycling rates is sustained only at lower EPS levels (below 2.79). At higher regulatory stringency levels, this effect diminishes and turns marginally negative.
Furthermore, the analysis highlighted the profound independent impact of institutional frameworks. The EPS index exhibited the largest direct effects across both models, drastically reducing waste generation (coefficient: -12.38) and substantially boosting recycling rates (coefficient: +7.38). This provides robust empirical validation for the Porter Hypothesis, demonstrating that rigorous environmental policies actively stimulate systemic efficiency and circularity rather than merely acting as economic constraints. Additional findings indicated that higher energy consumption correlates with increased waste and decreased recycling, while GDP per capita growth reflects absolute decoupling from waste generation, aligning with the Environmental Kuznets Curve. Interestingly, higher income levels negatively impacted recycling rates, a phenomenon theorized by Becker’s allocation of time model, where the rising opportunity cost of time in affluent societies deters household participation in time-consuming waste sorting.
Conclusion
The findings unequivocally demonstrate that the environmental impact of e-commerce is not absolute, but deeply conditional upon the prevailing institutional and regulatory landscape. While digital commerce acts as a potent catalyst for waste reduction and recycling in environments with nascent environmental policies, its comparative ecological benefits are largely neutralized under highly stringent regulatory regimes, where the rebound effects of packaging and fragmented delivery overshadow operational efficiencies. Therefore, policymakers must not view the transition to a digital economy as an automatic substitute for robust environmental governance.
To harness the synergies between the twin transitions of digitalization and sustainability, several strategic interventions are recommended. For advanced economies like the EU, transitioning from prescriptive regulations to smart, incentive-based frameworks is crucial. Policymakers should provide tax incentives for platforms adopting fully circular packaging and enforce Extended Producer Responsibility (EPR) mandates rigorously. The implementation of the Digital Product Passport (ESPR) is essential to eliminate the "free-rider" problem currently exploited by cross-border e-commerce sellers, ensuring full accountability across the global value chain. Moreover, urban planning should leverage population density by integrating municipal waste management systems with e-commerce delivery networks, transforming reverse logistics into a cost-effective, high-yield channel for recovering secondary raw materials. In developing regulatory environments, such as Iran, authorities should preemptively capitalize on the structural efficiencies of growing digital platforms to bypass traditional linear models, while concurrently institutionalizing EPR frameworks for dominant tech companies before a "throwaway society" culture becomes deeply entrenched.