TPRC 2024: The 52nd Research Conference on Communications, Information and Inter

Link to Paper

Abstract:
Background
Twin transitions, involving digital and sustainable transformations of contemporary societies and economies, have emerged as top priorities in the European Union, and important policy objectives of governments across the globe. For the business community and the economy more broadly, understanding the factors which influence the adoption of digital technologies by firms and linkages of these with sustainable business practices is important to inform policies aimed at fostering these twin transitions. This research paper generates new knowledge on linkages between digitalisation and the adoption of ICT-related sustainability practices across firms, sectors and space. Using firm-level data from Ireland, it examines three research questions:
1. Which firms are more likely to adopt ICT-related sustainability measures?
2. Does digitalisation affect the propensity of firms to adopt ICT-related sustainability measures?
3. What other factors influence the adoption of ICT-related sustainability practices across firms?
In terms of the existing body of evidence on these themes, the research provided in this paper presents a first analysis of linkages between digitalisation and the adoption of ICT-related sustainability practices across firms, sectors and space. It examines the usage of a range of advanced digital technologies such as Artificial Intelligence (AI), Internet of Things (IoT) and digital tools such as cloud computing services and software for sharing information within firm by firms in Ireland, and how these in turn relate to the adoption of different ICT-related environmental sustainability measures. The models employed estimate the propensity of firms to adopt a range of ICT-related sustainability measures as a function of digitalisation and other firm capabilities as well as external factors that influence the usage of digital technologies and the adoption of environmental sustainability measures, including for example competition from other firms as proxied by market share and exporting, and skills levels in firms.

Methods
The work is informed by a novel dataset which merges five micro-datasets for the first time. This includes the E-commerce and ICT Surveys for Irish-based enterprises conducted by Ireland’s Central Statistics Office for 2021 and 2022. These data were then merged with information on the same firms from the Census of Industrial Production, the Annual Services Inquiry, and the Business Register. Upon merging all data, and data cleaning, the final sample consists of 846 firms.
We examine the propensity of firms to adopt the following five ICT-related sustainability measures in 2022, whether the:
1. Firm considered the environmental impact of ICT services or ICT equipment
2. Firm employed measures to reduce the amount of paper used for printing and copying
3. Firm employed measures to reduce the energy consumption of ICT equipment
4. Firm disposed of ICT equipment in electronic waste collection or recycling
5. Firm reported keeping ICT equipment within the firm.
These dependent variables are modelled in separate probit regression models as a function of the usage of a range of digital technologies adopted as recorded in 2021, including, whether the firm reported:
1. Being a digital technology user: which is a dummy variable assigned a value of one whether the firm reported using at least one digital technology (Artificial Intelligence, Internet of Things) or a digital tool (cloud computing services, software for sharing information electronically with the firm); 0 otherwise
2. An AI user: firm reported use of AI
3. A IoT user: firm reported use of IoT
4. A Cloud computing user: firm reported use of cloud computing services
5. A User of software for sharing information electronically within firm: firm reported use of software for sharing information electronically within the firm; and,
6. The total number of digital technologies used by a firm
The regression models also encompassed a range of control variables captured in 2020, which may be associated with the usage of digital technologies and the adoption of ICT-related sustainability measures, including: market share, log of wage of employee, whether the firm had employed ICT specialists, whether the firm had an ultra-fast internet connection (100 megabytes per second), the log of investment in intangible assets (e.g. R&D), whether the firm was an exporter and whether they were an importer. The models also controlled for sector- and region-specific fixed effects. Unobserved omitted variables associated with the adoption of digital technologies and ICT-related sustainability measures such as managerial quality were assumed to be included in an error term.

Results
Descriptive statistics revealed differences in adoption of ICT-related sustainability factors by firm size, sector and region. The most adopted measures were the disposing of or recycling ICT equipment, while the least adopted measures were measures to reduce the energy consumption of ICT equipment. Regression modelling results demonstrated that the propensity of firms to consider the environmental impact of ICT services and ICT equipment was positively associated with the number of digital technologies used. The findings show that employing AI technologies, having a higher market share, having employees with higher skills, and employing ICT specialists was associated with implementing measures to reduce the consumption of paper for printing and copying. The propensity of firms to adopt measures to reduce the energy consumption of ICT equipment was positively associated with using AI and with firm size.
Turning to results on e-Waste management, the propensity of firms to dispose of or recycle ICT equipment was positively associated with the usage of software for sharing information electronically, employing ICT specialists and investing in intangible assets. In contrast, the propensity of firms to keep ICT equipment within the firm was negatively associated with investment in intangible assets.

Conclusions
The results of this paper add to the existing evidence suggesting that that the adoption of digital technologies and sustainability practices can be complementary strategic business objectives. The research provides novel evidence on the complementarity between the adoption of advanced digital technologies such as AI and IoT and other digital technologies, and the adoption of ICT-related sustainability measures. In terms of policy implications, the evidence suggests that where a wider adoption of ICT-related sustainability measures is desirable, targeted policy measures to incentivize and enable more firms, particularly small and medium sized enterprises, to adopt digital technologies and ICT-related sustainability measures could help foster the twin digital and green transitions. Policies to improve the adoption of sustainability measures by firms may also consider more broad firm factors including the upskilling of employees and employing ICT specialists, investing in research and development and other intangible assets such as computer software and organisational capital.
We note that the cross-sectional nature of the merged dataset makes causal relationships difficult to identify in this research. Furthermore, the data also only include firms with 20 or more persons engaged in the business, which limits the generalisability of the findings for very small or micro enterprises. Moreover, the results rely on self-reported information from firms which may be subject to measurement error and thus could lead to bias in estimations. For example, firms may tend to report greater adoption of sustainability practices to appear ‘greener’.
How digitalisation and sustainability feature in firm decision making and operations is still an emerging academic topic. The use of longitudinal data in future research, and data from other jurisdictions are required to und

Link to paper

Abstract:
The environmental impact of large tech companies has come under intense scrutiny, particularly as the COVID-19 pandemic and the frenzy about Generative AI have driven increased demand for ICT infrastructures and services. Unsurprisingly, calls for more transparency and the disclosure of sustainability-related metrics by these companies have grown over time. This paper explores the sustainability reporting strategies of large tech firms, with a focus on decarbonization and data centers. We examine critical areas such as greenhouse gas emissions, renewable energy usage, and energy efficiency through a longitudinal comparative analysis of sustainability reporting activities from 2003 to 2023. Our sample includes nine leading tech companies from the US (Akamai, Amazon, Apple, Facebook/Meta, Google/Alphabet, Microsoft) and China (Alibaba, Baidu, Tencent). The analysis highlights a convergence among these companies towards publishing dedicated sustainability reports, providing more comprehensive and nuanced insights into supply chains and data centers, and emphasizing carbon neutrality and 100% renewable energy usage. A notable trend is the shift in focus from energy efficiency to renewable energy supply and from operational emissions to value chain emissions. Based on a critical discussion, we offer valuable insights for different stakeholders.

Link to paper
Abstract:
The introduction of 5G technology has changed the mobile broadband landscape, yielding faster speeds, lower latencies, and potentially more widespread coverage. In this paper, we study the evolution of US cellular technology, with a focus on 5G, from 2021 to 2023 through the lens of Ookla Speedtest performance and cellular infrastructure deployment. Specifically, we analyze how US cellular network coverage and performance evolved during this period, characterizing improvements at multiple geographic granularities, examining differences between urban and rural performance, and evaluating the availability and performance of “Fast 5G.” Then, we investigate the impact of cellular infrastructure, studying the relationship between infrastructure deployment density and network performance and the growth in infrastructure during this time period. Our findings show that, broadly, mobile network performance improved, though the improvement in some states and regions was far greater than in others. For instance, some states show an increase in download speeds of over 200%, while other states show little to no improvement. Cellular deployment density also grew during this period, is approximately 15 to 40 times higher in urban areas than in rural areas, and is strongly correlated with population density. We find that higher cellular deployment density is generally associated with improved network performance; however, the growth in deployment density does not always align with performance gains, as evidenced by weak correlations between increases in density and improvements in performance metrics. We conclude with recommendations about the need for more granular data about cellular technology infrastructure, deployment dates and location data to better inform policymakers about targeted investments in additional cellular infrastructure.