The ongoing digitalisation brings about significant societal and economic changes, in some instances profoundly changing the ways we live and work. This trend also induces various and potentially far-reaching energetic and environmental effects. The direct energy and environmental footprint of digitalisation, for example, is a growing source of concern, in particular in the context of the latest developments within artificial intelligence (AI). Digitalisation can, however, also induce indirect benefits in other domains through various mechanisms such as efficiency gains, dematerialisation, better coordination, or simulation and forecasting.
Unfortunately, increasing evidence shows that these very mechanisms, often induce environmentally detrimental indirect effects as well. These undesired side-effects appear because of the decreased prices, saved time, or increased convenience brought about by the positive developments mentioned above. They can reduce or even reverse the initial positive effect, and are often described in economics under the umbrella term of rebound effects. Unlike its direct footprint and its positive indirect effects, the rebound effects of digitalisation have received little attention so far.
The current study sheds light on the rebound effects that occur in two domains that have recently been profoundly affected by digitalisation, especially since the pandemics: shopping and working. Enabled by digitalisation and accelerated by the pandemics, e-commerce and teleworking have substantially altered the way we shop and work. E-commerce and teleworking both reduce the need for in-person trips to stores and work, and thus have the potential to reduce energy consumption and reduce emissions of greenhouse gases and pollutants.
Building on comprehensive literature reviews in both domains, this study shows, however, that both domains are also characterised by strong rebound effects. For e-commerce, rebound effects are triggered by the additional ICT systems needed, the additional packaging usually deployed, the (typically van-based) last-mile delivery, failed deliveries followed by re-deliveries, and the returns, which are substantially higher for online than for traditional in-store shopping. For teleworking, rebound effects occur due additional ICT needed to support the process, the additional domestic energy consumption, the more frequent non-commute trips, and the longer commute that teleworkers are willing to perform on the days when they do commute to work.
Overall, both rebound effects are similarly-sized, amounting to a few kilograms of CO2 per online shopping instance and teleworking day, respectively. Aggregating these effects over the dozens of online shopping processes and teleworking days per year results in an estimated median of 111 kg CO2/year rebound effects of e-commerce per average Swiss household, and 91 kg CO2/year rebound effects of teleworking per Swiss inhabitant, which – for Switzerland – represent 1.02% and 1.82% of the per-capita greenhouse gas emissions, respectively.
In both cases, the additionally required ICT infrastructure induces only a marginal effect. Except the additional domestic energy required for teleworking, most other types of rebound effects for both e-commerce and teleworking stem from additionally generated traffic. Substituting environmentally friendly transportation means for traditional ones thus helps mitigating the rebound effects in both cases. Such substitutions can result from the general transportation infrastructure and culture, public policies, company policies, or individual choices.
Among the two domains, the rebound of teleworking seems higher, but at the same time more clearly defined and limited in scope, and thus relatively straightforward to understand and – to the extent possible – mitigate. For e-commerce, however, there are further, quite subtle, hard to grasp and to control sources of rebound: the consumption-inducing effect of increased convenience and new marketing channels, the intercontinental, often airplane-based deliveries and returns enabled by new distribution platforms, and the high destruction rate of returned items, resulting both from the high return rate of e-commerce and the non-saleability of quickly outdated goods such as fast fashion and electronics. More research is needed to better understand these effects and to devise policy measures for their mitigation, as well as to better conceptually describe the indirect effects of digitalisation.