Measuring Science and Innovation for Sustainable Growth

Foreword

The OECD “Measuring” monographs on Science, Technology and Innovation (STI) are internationally recognised vehicles for communicating analysis of key and novel aspects of science and innovation. They address gaps in the provision and use of STI indicators and serve as reference documents to drive the measurement and analysis agenda for better, evidence-based policies, both at OECD and across its members and partners. Past “Measuring” publications have informed the OECD Innovation Strategy and the Organisation’s monitoring and recommendations on digital transformation, contributing to mainstreaming these policies across government and providing new foundations for measurement and analysis.

Following this model, this “Measuring” publication aims to articulate measurement concepts and provide new statistical insights on the multidimensional contribution of STI to sustainable growth in the context of key energy and green transitions. This goal is also supported by the Declaration on Transformative Science, Technology and Innovation Policies for a Sustainable and Inclusive Future, which was adopted at the meeting of the OECD Committee for Scientific and Technological Policy (CSTP) at the ministerial level in April 2024. One of this declaration’s four pillars focused on “Strengthening the evidence base for STI strategies and policy making”, specifically calling on the “OECD, through the CSTP and in collaboration with other relevant OECD committees” to “provide international statistics, data resources and new evidence on STI systems and policies and their impacts, notably regarding the Sustainable Development Goals and the just and green transitions.” The publication is the result of a project approved by the OECD Committee for Scientific and Technological Policy (CSTP) and included in its Programme of Work and Budget (PWB) 2023-2024.

This report, informed by preparatory work and engagement across OECD bodies and with the International Energy Agency, aims to serve as a reference document for policymakers, statisticians, researchers and the wider public, bringing together a diverse range of data sources and outputs to give a comprehensive and international overview of the channels through which science and innovation systems contribute to sustainable growth. The additional overarching aim is to provide a measurement blueprint to lay the ground for a forward-looking international agenda for continuous improvement in the quality of measurement of sustainability-relevant science and innovation.

Three objectives underpin the overarching aims:

• Map and render operational concepts and definitions of “green” science and innovation.

• Provide a comprehensive and interconnected (statistical and data-based) view of the STI system’s role as an enabler of the green transition, drawing on available indicators.

• Identify knowledge gaps and limitations of existing indicators, showcase experiences demonstrating the feasibility of measurement, and introduce new measurement experiments.

Acknowledgements

The report is the result of collective effort coordinated by the Secretariat to the OECD Working Party of National Experts on Science and Technology Indicators (NESTI) at the Science and Technology Policy Division (STP) in the OECD Directorate for Science, Technology, and Innovation (DSTI), with Petra Kelly and Fernando Galindo-Rueda as project managers under the guidance of Alessandra Colecchia as Head of the STP division.

This work would not have been possible without the substantive contributions of several DSTI colleagues, including Leonidas Aristodemou, Silvia Appelt, Kaijie Ou Huang, Guillaume Kpodar, Lorine Labrue, Adrian Leung, Kuniko Matsumoto, Daniel Sánchez-Serra, Blandine Serve, Brigitte Van Beuzekom and Fabien Verger in the STP division; Marius Berger, Antoine Dechezleprêtre, Hélène Dernis, Antton Haramboure, Nuala Mulligan and Alžběta Vítková from the Productivity, Innovation and Entrepreneurship division; Brigitte Acoca and Jan Tscheke from the Digital and green consumers unit; and Lenka Wildnerová in the Steel unit.

Outside DSTI, the publication coordinators would like to acknowledge the contributions by Alberto Agnelli (ENV/EEI), Olof Bystrom (ENV/EEI), Damien Dussaux (ENV/EEI), Katherine Hassett (ENV/EEI), Rose Mba Mébiame (ENV/EEI), Mauro Migotto (ENV/EPI) in the Environment Directorate; Tobias Kruse in the Economics Department, Carlo Menon from the Centre for Entrepreneurship, SMEs, Regions and Cities, Carlos Hinojosa from the Evaluation and Internal Audit (EVIA) unit and Stefano Contratto from the Directorate for Communications. EVIA and the Directorate for Communications kindly facilitated access to Overton data.

In addition to OECD sources, the report relies on valuable contributions and comments from the International Energy Agency kindly provided by Oskaras Alsauskas, Simon Bennett, Elisabeth Connelly, Amrita Dasgupta, Shobhan Dhir, Roberta Quadrelli, Mathilde Huismans, Suzy Leprince, Teo Lombardo and Apostolos Petropoulos. Other organisations and individuals who have contributed to this report include the Joint Research Centre, the International Renewable Energy Agency, the International Monetary Fund, Our World In Data, Overton, Eurostat, the BBVA Foundation, the Fraunhofer Institute and Matěj Bajgar (Charles University).

The coordinators would also like to thank DSTI Director Jerry Sheehan and Deputy Director Jens Lundsgaard and their team, especially Alice Holt, Takako Kitahara and Martina Fattiboni Ferrara, for valuable comments and assistance in conducting consultations, as well as Sylvain Fraccola, Delphine Kadysz, Joe Loux, Sebastian Ordelheide, and Kyriakos Vogiatzis, who provided invaluable for communications and administrative guidance and support. The work has also benefited from STI policy- oriented comments by several STP colleagues, including Mario Cervantes, Charles McIvor, Michael Keenan, Philippe Larrue and Carthage Smith and from editorial review by Julie Harris.

The study was conducted as part of the Programme of Work and Budget 2023-24 of the Committee for Scientific and Technological Policy (CSTP), under the auspices of the Working Party of National Experts on Science and Technology Indicators (NESTI) and its multiple informal networks, including the R&D statistic experts network, which furnishes data on R&D by socio-economic objectives and provided several examples of cross-cutting energy and environmental R&D and the Innovation survey co-ordination group, which drove the collection of data on Innovation with Environmental Benefits. The report coordinators would also like to thank the members of NESTI’s Expert Group on Management and Analysis of Research and Innovation Administrative Data (MARIAD), who directly oversee the Fundstat work mapping R&D project funding and its potential relevance and impact.

Earlier versions of this report were presented for discussion at the 2024 NESTI official meeting. Selected results were also presented in the internal brown bag seminar on “Measuring Science and Innovation for Sustainable Growth” in March 2025, at the CSTP official meeting in April 2025 and at the official meeting of Working Party on Innovation and Technology Policy (TIP) in June 2025. The coordinators would like to express their gratitude to NESTI, CSTP and TIP delegates for their feedback.

Lastly, the coordinators would like to thank the ministers and senior officials participating in the breakout session, which took place as part of the OECD Science and Technology Policy Ministerial meeting, whose background paper provided the initial basis and motivation for Chapter 4.

Reader’s guide

Acronyms

Executive summary

The transformational power of science, technology and innovation (STI) is necessary to extend economic wellbeing while addressing major global environmental challenges such as climate change, pollution and biodiversity loss. To harness and direct this power effectively, however, policymakers require fit-for-purpose measures that capture the multifaceted contribution of science and innovation, the role played by different actors and the impact of relevant policies. Measuring Science and Innovation for Sustainable Growth provides new and comprehensive evidence on the role of science and innovation for resource and environmental sustainability. It draws on existing statistics and newly developed databases, indicators and analysis to map key trends and structural issues, comparing the capabilities of countries in this important domain to inform decision making. The following insights emerge from the gathered evidence:

Environmental innovation is transforming markets and making low-carbon energy and transport more affordable faster than previously thought possible.

• Cost reductions enabled by technological innovation have accelerated the uptake of new technologies across key sectors. In 2023, for instance, electricity from large-scale solar PV was 56% cheaper than from fossil fuels—having been over four times more expensive in 2010. The global electric car fleet reached nearly 58 million by the end of 2024, more than triple its size in 2021.

• In the median OECD country, over 40% of innovative firms introduced at least one environmentally beneficial innovation in the previous three years.

• Environmental technology startups have attracted 17% of total venture capital (VC) funding, and account for a growing share of national startup populations in three out of four surveyed countries.

Science underpins energy and environmental innovation through critical, often underappreciated, channels.

• New indicators show that nearly 28% of scientific publications are relevant to energy and environmental societal goals.

• International collaboration is significantly more prevalent in energy and environmental science than in other fields.

• Climate change mitigation and adaptation patent filings build more directly on scientific knowledge than high-carbon ones, citing nearly six times as many peer-reviewed articles. Key cited disciplines include engineering, chemistry, and materials science, with computer science also playing a greater role in low-carbon than in high-carbon patents.

• Start-ups in energy and sustainability are the second most likely to be founded by PhD holders (16%) - underscoring science’s role in entrepreneurial activity.

OECD countries no longer lead on science and innovation addressing energy and environmental goals.

• China is now the world’s largest contributor to scientific publications in energy journals and accounts for 40% of the 10% most cited environment- and energy-related publications.

• China’s high-quality patent filings in environmental technologies grew more than sixfold between 2010 and 2020, while growth in the United States and the EU remained below 20%. Nearly a quarter of China’s international patent filings now relate to environmental outcomes, compared to just 10% in the United States.

• The share of venture capital channeled into environment-related start-ups grew by 25 percentage points in China between 2010 and 2022, compared to just 6 percentage points in the EU and stagnation in the wider OECD.

• China also dominates clean technology manufacturing and leads in exports of environmental goods, having gone from laggard to leader over three decades.

Public support for energy and environmental research and innovation has been sluggish for decades but has been picking up recently.

• Technology-specific support for low-carbon innovation declined across the OECD between 2012 and 2019, even as environmental policy stringency increased.

• Between the early 1980s and 2006, government budgets for R&D dedicated to energy and the environment have remained largely flat in real terms. Relevant R&D budgets in the OECD area grew by 29% in 2023 and stood at 7% on average that year. Less than 5% is dedicated to low-carbon energy research, development and demonstration (RD&D).

• Recovery spending channeled into innovation with environmental objectives post-COVID was heavily weighted toward technology deployment, with just 6.5% allocated to RD&D.

• Analysis of R&D funding portfolios for governmental funding bodies shows that energy and environmental goals accounted for approximately 20% of funding in 2023, representing nearly USD 40 billion.

Directed R&D support appears more effective in supporting environment-related innovation than neutral instruments.

• New analysis shows that direct public funding for business R&D is positively associated with the uptake of innovations with environmental benefits, particularly those that reduce CO₂ emissions, pollution, or resource intensity. Environment-related patenting is also more common among firms receiving direct support.

• R&D tax incentives, by contrast, show no clear effect in EU countries on environmental outcomes or the composition of private R&D. Firms cite market demand, reputation, regulations, and cost pressures as key motivators for pursuing environmental innovation.

Public opinion supports science and innovation for the green transition.

• Environmental concerns remain high on the public agenda, especially in the EU, even as safety and economic concerns loom large. The percentage of respondents who agree or strongly agree that environmental issues will be resolved primarily through technological progress is the highest in Israel (62%), followed by Sweden (48%) and United States (47%).

• Trust in climate scientists is strong across the world and there is support for climate action - in particular in the form of low-carbon technology subsidies.

• Carbon taxes are more acceptable when linked to clean technology investments, indicating public appetite for coherent, science-informed policy solutions.

Focused, concerted action is needed to generate evidence on science and innovation’s contribution to environmental goals and to better inform policy.

• Major evidence gaps hamper decision making at the intersection of science, innovation, and sustainable growth. Governments can respond by adopting a three-pillar measurement agenda:

• Pillar 1: Build the foundations Leverage diverse data sources and methods with multi-purpose planning; align measurement efforts toward environmental science and innovation; and foster shared ownership and partnerships for effective data use.

• Pillar 2: Address evidence gaps Capture the broad contributions of upstream research; track green technology adoption and the reach of environmental innovation beyond business; monitor public support beyond R&D; assess systemic STI roles in the economy and environment; and account for STI’s own environmental footprint.

• Pillar 3: Measure impacts Develop systems of indicators tailored to policy needs and embed evaluation in the design and implementation of innovation policy.

1. The case for measuring science and innovation for sustainable growth

Abstract

This introductory chapter sets out the rationale for this publication’s in-depth look at measuring the contribution of science and innovation to sustainable growth. It outlines the main mechanisms through which science and innovation may contribute to sustainable transitions and provides the basis for the overall structure of the publication. It makes the case that measuring science and innovation’s contribution to sustainable growth plays a vital role in informing policy and the broader public debate. The chapter also lays out the main challenges for tracing the impact of innovation on sustainable growth. Finally, it describes the approach followed in the remainder of this publication to develop indicators and address gaps.

Science and innovation for sustainable growth

The world today faces several headwinds against preserving and expanding economic opportunities for current and future generations. These include natural resource and environmental pressures that, if left unaddressed, undermine not only society’s ability to seize new opportunities but also risk compromising well-being standards. Governments have taken steps to implement energy, climate and environmental policies, working to achieve a mutually agreed goals to improve natural resource and environmental sustainability. However, rising geopolitical tensions, supply chain disruptions, inflationary pressures, and a growing emphasis on national security have contributing towards reshaping policy priorities. In this context of complex interdependencies across policy objectives, governments face multiple imperatives and difficult choices in pursuing sustainable economic growth (Box 1.1).

Box 1.1. What is sustainable economic growth?

Paraphrasing the definition of sustainable development originally introduced by the Brundtland Commission (WCED, 1987[1]), sustainable economic growth refers to economic growth that meets the needs of the present without compromising the ability of future generations to meet their own needs. This concept of sustainability highlights the key role of natural assets in shaping the planet’s potential to sustain economic growth, i.e. the increase in living standards as measured by the value of goods and services produced in an economy.

Similarly, the Green Growth Strategy and framework adopted by the OECD recognised natural capital as a key factor of production and its central role in preserving and enhancing the well-being of current and future generations (OECD, 2011[2]).

In this challenging context, leveraging the full potential of science and innovation becomes critical, and so does the evidence base upon which such choices are based. Indeed, scientific and technological developments have in the past helped humankind address seemingly intractable resource constraints that imposed subsistence-level livelihoods.

The example provided by the demonstrable success of the Montreal Protocol in addressing ozone layer depletion illustrates the multiple channels through which science and innovation can underpin solutions to complex and global environmental problems. These provide a guide to measurement approach in this document. Scientific understanding of the thinning of the ozone layer, its causes and impact on human health helped not only raise awareness of the issues but also galvanised public opinion and helped forge an international consensus on the case for action. However, science-based awareness was a necessary but not sufficient condition for transformation. Action was also enabled by the rapid technology-based innovation in developing and implementing alternatives to ozone-depleting substances, which allowed for the timely substitution of the most damaging pollutants in the marketplace (Whitesides, 2020[3]; Gonzalez, Taddonio and Sherman, 2015[4]).

This experience, while not necessarily representative of how other pressing natural resource and environmental challenges can be addressed, highlights the combined potential role of science, technology and innovation (or “science and innovation” or STI hereafter for brevity) in enabling sustainable growth, spelling out the roles of scientific knowledge, technology development, market adoption and diffusion and society readiness and policies. These functions effectively articulate the structure of this publication.

Reflecting this potential and the systemic features of science and innovation, the OECD Agenda for Transformative Science, Technology and Innovation Policies was endorsed in April 2024 by science and innovation ministers from OECD and several other countries. This document argues that sustainable economic growth calls for substantive and multifaceted transformations that call upon the comprehensive mobilisation of the science and innovation system (OECD, 2024[5]).

The road ahead towards achieving sustainable growth objectives remains long as the adoption of the currently best available technologies and practices, while necessary, is far from sufficient. For example, the International Energy Agency (IEA) estimated that in order to limit the global temperature rise to 1.5 °C with at least a 50% probability without sacrificing economic growth, 35% of the required greenhouse gas reductions in 2050 would need to come from technologies that are not yet on the market (IEA, 2021[6]). Technologies that have yet to prove their potential and reach the market are particularly critical for certain sectors where pollution is currently difficult to abate, such as steel, cement, or maritime transport (OECD, 2025[7]).

There are justifiable concerns that the uptake of “net zero” and other sustainability-related commitments and the implementation of increasingly stringent policies to achieve them may, without complementary policies, introduce excessive costs on businesses (Stock, 2022[8]) and lead to a loss of economic activity in those jurisdictions that take a lead, for instance through the so-called “carbon leakage” in the context of climate change regulation (Dechezleprêtre et al., 2022[9]; Fowlie and Reguant, 2018[10]). Technological advances are therefore key to mitigating the risk of negative economic and social consequences that arise from policies, such as carbon pricing mechanisms, that impose a direct cost on certain sectors and communities with limited capability to adjust or be compensated. For instance, strategic investment in research and development (R&D) can reduce the carbon price that is needed to achieve a given level of emission reductions, thus sheltering industries and their induced employment from possible detrimental consequences of high carbon prices (Acemoglu et al., 2012[11]). By the same token, innovations that make low-carbon technologies more affordable can help mitigate passed-through price impacts of environmental policies. Without innovation, “net zero” targets can only be achieved at a much higher cost (Creutzig et al., 2023[12]).

Research breakthroughs underpin much of the progress towards technological discoveries and their eventual adoption, ultimately enabling the expansion of economic opportunities for production and consumption while reducing environmental harm. This encompasses not only applied research, which as defined in the OECD Frascati Manual is “original investigation undertaken in order to acquire new knowledge, directed primarily towards specific practical aims or objectives”, such as better environmental performance and cost-effectiveness; but also, basic research, which is “experimental or theoretical work undertaken primarily to acquire new knowledge of the underlying foundation of phenomena and observable facts” (OECD, 2015[13]). Curiosity-driven research is a major factor of serendipity in scientific breakthroughs that open pathways to solutions previously deemed unworkable.

As illustrated in the previous example about the ozone layer, science is also critical for generating reliable knowledge of the natural environment with its complex array of feedback loops and tipping points. It is an essential prerequisite for understanding the rate and scale of environmental change, including the factors driving it. Last but not least, science across all domains can play an additional key role in informing public policy and private behaviour by enabling an enhanced and trustworthy understanding of change and performance of environmental, technological and socio-economic systems.1

The measurement imperative

Measuring science and innovation for sustainable growth matters

It is important to develop an accurate and realistic appreciation of uncertainty about the potential and limitations of science and innovation to affect desired changes.2 The outcomes of different research and innovation activities result in a wide range of possible short- and longer-term environmental and economic outcomes. Exploring questions such as the role of policy and its impacts requires careful study of how innovation systems work. If science and innovation policies are going to have the intended effect of enabling desired energy and environmental transitions, they need to be supported by effective measurement and analysis of a qualitative and quantitative nature. Timely, reliable and comparable data and information on the role of STI systems in this area, and its policies, inform steps to support desired transitions and shed light on progress and outcomes, including impacts on people and communities.

There is a strong global measurement culture underpinning environmental monitoring and policy. The OECD provides foundational data and metrics to assess country progress towards several environmental goals and distil the main implications for economic collaboration and economic development policies (OECD, 2012[14]; OECD, forthcoming[15]). Likewise, the IEA monitors the global energy system, providing a comprehensive view of its past evolution and future outlook (IEA, 2024[16]). The OECD Council Recommendation on Environmental Information and Reporting (OECD, 2022[17]) recommends that “adherents take a comprehensive approach to improving environmental information and reporting, as well as information systems and measurement frameworks.” It calls for improvements in “scientific knowledge, information, statistics, accounts and indicators on the environment and sustainable development, in order to contribute to the evaluation of: a) the state of the environment; b) activities that affect or are likely to affect the environment; c) policies, plans, actions and programmes that affect or are likely to affect the environment; d) environment policies themselves.” This mandate not only explicitly calls out the role of scientific knowledge and statistics but also provides the basis for considering, from a measurement perspective, science and innovation as human activities that affect the environment.

The challenge of measuring science and innovation for sustainable growth

Evidence gaps and uncertainties about the ability of economic and social systems to support sustainable growth call for specific measurement and scientific research efforts in this area. Within the objective of improving understanding of activities that affect or are likely to affect natural resources and the environment, understanding the role of science and innovation systems is potentially one of the most challenging, but by the same token, one for which there is significant potential for improvement and positive impact for decision makers. Innovation and environmental performance can be particularly challenging and “evasive” objects of measurement.

Science and innovation possess unique features as a domain for policy analysis owing to the intangible nature of knowledge and how it is created and diffused, which leaves few unambiguous and reliably measurable traces (OECD, 2018[18]). The outcomes of innovation efforts can be highly uncertain and valuable to those concerned. This makes it difficult not only to anticipate what comes out of them but can also create incentives to conceal or exaggerate claims, depending on what is at stake. Knowledge about the wide range of different STI activities can be a source of strategic comparative advantage for individuals and organisations that compete for resources, market share and hegemony.

Furthermore, the empirical study of innovation and innovation policy faces the challenge of seeking to measure how activities that are themselves difficult to measure affect other outcomes that are also difficult to measure. This applies when attempting to identify and demonstrate the link between STI activities and potential or actual environmental sustainability outcomes that can also be uncertain or far in the future.

Characterising environmental sustainability is an equally challenging empirical question when it comes to economic activities in general and science and innovation in particular. As environmental sustainability is broadly perceived as positive, and several policies promote it actively, there are incentives for overstating “green” claims, i.e. what is commonly known as “greenwashing” although in recent years there is increasing evidence of the opposite behaviour, also referred to as “greenhushing”. This can distort measurement of the relevance and impact of all economic activities towards environmental sustainability objectives (ESOs), including science and innovation (Box 1.2).

Box 1.2. “Greenwashing”, “greenhushing” and their implications for measurement and analysis

Goodhart’s Law states that “when a measure becomes a target, it ceases to be a good measure”. This maxim captures the predicament that once a metric is considered an indicator of success or effective compliance, there is an incentive to manipulate it. When formulated with respect to environmental credentials of products and services, this phenomenon is also often known as “greenwashing” (Gatti, Seele and Rademacher, 2019[19]; OECD, 2025[20]). More recently evidence of “greenhushing” - a practice which refers to organisations deliberatively under-communicating or remaining silent about their environmental efforts - has also started to emerge (Ginder, Kwon and Byun, 2019[21]).

Greenwashing and greenhushing can affect surveys as respondents may seek to show they are behaving in a manner they interpret as desirable. While confidentiality and other survey protocol adjustments can be implemented counter bias of all types (Bloom and Van Reenen, 2007[22]; Yong et al., 2024[23]), these may not be fully effective in all contexts. Behaviours such as “satisficing”, where the respondents select the first acceptable answer or drop out of surveys to reduce further burdens, may be further sources of measurement error, though not necessarily only affecting sustainability-relevant science and innovation indicators. Greenwashing, greenhushing and other types of bias might also distort indicators based on data other than surveys, such as indicators based on corporate disclosures and descriptions of R&D projects and firms or data reported for administrative purposes. Information is revealed in those instances to meet compliance or strategic aims and may not necessarily be accurate.

Building awareness and developing new techniques to attenuate possible biases in suitable datasets (e.g. text-rich opportunity data) offers a potential avenue for developing a new generation of more robust indicators. Policymakers need to work with indicator experts to put in place robust, hard-to-cheat systems, approaching the challenge strategically. As the stakes (economic, reputational, environmental) keep growing, the larger the risk that biased information will displace quality data and analysis. A perspective of “policies for good data” should try to ensure, when monitoring conditions are imperfect, that reporting and measurement are linked to small-stakes outcomes that induce truthful reporting; otherwise, they are likely to be the object of major distortions.

Source: Authors’ own elaboration.

Figure 1.1 sums up the measurement challenge by spelling out the intrinsic challenges of innovation measurement, the uncertain and multidimensional nature of environmental sustainability indicators and the specific challenges of inferring relevance and impacts from limited available data sources. Broadly speaking, there are two main approaches for establishing connections in the data between STI and environmental sustainability.

• A direct approach is to rely on what data providers themselves report to be the case against specific questions and reporting requirements on green outcomes, which might map on to ESO or related “green” taxonomies. The data providers may be the actors themselves or third parties accrediting some particular aspect, such as a patent examiner confirming the domain of a patent application and its fulfilment of patentability requirements.

• A derived approach is to draw inferences from the reported information, e.g. when the party responsible for measurement deploys judgements and tools aimed at characterising the STI activity on which it has information. This might be the case of an analyst using artificial intelligence (AI) tools to infer patterns about the relevance of an R&D project or a public procurement contract description to some environmental outcome. It may also apply when a statistician connects data on the environmental performance of the actor against what it reports in terms of STI activity.

Figure 1.1. Measurement challenges connecting science and innovation with environmental sustainability objectives and outcomes

Source: Authors’ own elaboration.

Measurement of relevance, e.g. through application and goal-oriented perspectives, can be biased against upstream knowledge activities for which the link may not be explicit or deterministic. Such a goal-oriented approach may, for example, contribute to understating the role of basic science and its contribution to sustainable growth since the information and statistical categories that describe activities upstream of STI are formulated in terms that are distant from application goals. A comprehensive study on science and innovation for sustainable growth needs to account for these indirect but foundational contributions.

Conversely, many indicators on STI, especially those relating to policy levers, tend to focus on available indicators on support for R&D and, as a result, miss out on support provided for demonstration and deployment activities closer to the market and final users, so there can also be biases against capturing several downstream activities.

Against this backdrop, the evidence required to inform policy for sustainable growth cannot rely solely on descriptive measurement. In order to empirically assess impacts, it also has to incorporate modelling and counterfactual analysis, which requires an effective data cycle in which ex ante and ex post evaluation build on each other and draw upon reliable measurements.

In the green transition challenge, policymakers’ main motivation for evidence is to be able to assess options lying ahead to drive consensus on the course of action and have the relevant means to reassess policies. To achieve this, effective governance also requires measurement for accountability to attribute merit and rewards. The options that policymakers need to consider entail complex trade-offs and a multiplicity of actors, so tools are required to bring disparate data sources together.

One key challenge for building evidence in this area is the limited capacity to bring together data on innovation inputs with data about material flows that matter for energy and sustainability transitions. While progressively advancing at a general level, the statistical data-linking agenda has not been moving as fast as required by the severity of the policy challenge and needs to go beyond the domain of economic statistics. This requires more effective policy co-ordination and regulation to ensure secure spaces in which data that may be deemed confidential can be safely processed and analysed to its full potential.

This publication’s approach and methods

Use and connect existing definitions and taxonomies

Working at the measurement interface between science and innovation and sustainable growth requires consistent, clear and unambiguous definitions and classifications. This helps build a reference framework and build towards developing and maintaining consistent and internationally comparable measurements over time to inform decision making. In this “Measuring” publication initiative, concepts and taxonomies relating to resource and environmental impact, as well as those relating to science and innovation, must be brought together.

A structure based on the depiction of the science and innovation system

This publication depicts a variety of indicators under a comprehensive narrative of potential impact and relevance pathways for science and innovation, as outlined in this introduction in its depiction of the science and innovation system. This covers impact pathways for knowledge generation and activities concerning its practical application, such as the adoption of innovation. It also covers specific indicators on the roles played by government policies, markets and society. These different elements provide the structure for the publication.

A broad-based view of resource and environmental sustainability and its taxonomies

This publication adopts a flexible view and gathers detailed information about the definitions used. It classifies relevant data and indicators according to the definitions and taxonomies3 they rely on, both along the research and innovation dimension and along the environmental dimension. It does not limit the scope of the enquiry only to specific environmental issues (e.g. climate change only).

There are relatively few instances of environmental taxonomies embedded into STI measurement. For example, in the area of measuring business innovation, international standardisation efforts that began in the early 1990s efforts have not yet converged into an internationally agreed definition and classification of energy or environment-related innovation. Some measurement experiences, such as those initially promoted by the European Community Innovation Survey, provide an initial basis for international comparisons.

Critical use of taxonomies for classification purposes

Existing standard classifications used to report the broad socio-economic objectives of government support for R&D help to identify funding for environmental and energy objectives. However, they do not provide a means to assess whether, for example, energy-focused support for R&D or block grant support for R&D in universities and major research institutes contribute to environmental sustainability.

This publication makes extensive use of methods developed through parallel, experimental work to monitor the relevance of science and R&D efforts towards the United Nations Sustainable Development Goals (SDGs) across some of its new experimental measurement work, an initiative first put in motion following the OECD Blue Sky Conference (OECD, 2018[18]). The use of classification frameworks for societal goals such as SDGs or the NABS (Nomenclature for the analysis and comparison of scientific programmes and budgets) in classifying socio-economic objectives highlights that relevance to an objective is not mutually exclusive of others. However, its operationalisation for quantification purposes that allow adding up to meaningful totals requires the use of apportioning or prioritisation procedures for the units of analysis classified. This work reveals the limitations and trade-offs of classification frameworks and their practical application to different purposes.

Data sources and methods for assessing relevance

Build on existing “designed” and “opportunity” data

A significant part of this work is based on a diverse body of existing indicators, rarely presented under a common structure. The publication assesses their strengths and limitations in terms of providing robust and unbiased measurement of a particular element of the STI system with respect to the green transition. Adjustments to these well-established indicators have been made wherever possible to maximise their relevance.

Measurement of science and innovation can be based on a diverse set of sources and draw upon multiple methods (Figure 1.2). Statistical surveys, especially those designed with the explicit purpose of providing representative and comparable statistics, are a primary source of information at the national level and a keystone of international benchmarking efforts. Surveys can allow respondents to self-declare green attributes of their STI activities and outputs within the space and other limitations of the survey vehicle and the respondent’s knowledge and incentives.

While surveys are well suited to answering certain types of questions and can elicit valuable information from a wide range of important stakeholders, including firms and households, their use to inform analysis of “STI” for sustainable growth is still limited and faces significant challenges that limit the ability to introduce new, more targeted, questions. In a complementary fashion, this publication also exploits opportunities provided by additional information sources generated for purposes other than producing statistics, e.g. for administrative or commercial purposes. Some of those are relatively well tested and understood, such as patent data, while others are still in the early stages of development, such as databases on R&D project awards, company financial reports and others that are being increasingly exploited to deliver policy-relevant insights. The publication thus dedicates considerable effort to laying out what available administrative sources may be best suited for and where the main limitations lie.

Figure 1.2. Main types of data sources on science and innovation

Source: Authors’ own elaboration.

Experiment in assessing environmental sustainability in relevance in STI data

Many internationally comparable indicators and measures already exist and reveal policy-relevant information regarding the inputs or outputs at various points of the science and innovation chain, as well as regarding eventual outcomes. The aim of the publication is, however, also to go beyond these existing indicators and develop new ones that are thus far not available in OECD or other broad intergovernmental fora. To that end, the publication exploits relatively new data sources and methods that may offer promising avenues for gathering additional insights to fill known measurement gaps.

Examples of measurement experiments based on the application of advanced methods include the use of:

• Natural Language Programming (NLP) AI methods for identification and classification of text descriptions of STI activities

• combined use of survey and administrative data

• data linking of disparate data sources to connect STI inputs and outputs.

Showcase experiences demonstrating potential

Where appropriate, when international indicators are not readily available and not possible to derive from opportunity data, the publication will rely on single-country deep dive analyses and case studies, both quantitative and, where appropriate, qualitative. Showcasing novel approaches is vital to enriching understanding of the multitude of ways in which STI underpin the green transition.

Identify gaps and recommend possible future steps

This publication aims, above all, to promote action towards improving understanding how science and innovation contribute to sustainable growth. Available indicators provide only a partial picture that barely scratches the complexity of issues and uncertainties at stake. This publication’s final and concluding chapter sets out some main takeaways and proposals for a measurement agenda in this vital area.

References

[11] Acemoglu, D. et al. (2012), “The environment and directed technical change”, American Economic Review, Vol. 102/1, pp. 131-166, https://doi.org/10.1257/aer.102.1.131.

[22] Bloom, N. and J. Van Reenen (2007), “Measuring and Explaining Management Practices Across Firms and Countries”, The Quarterly Journal of Economics, Vol. 122/4, pp. 1351-1408, https://doi.org/10.1162/qjec.2007.122.4.1351.

[12] Creutzig, F. et al. (2023), “Technological innovation enables low cost climate change mitigation”, Energy Research and Social Science, Vol. 105, p. 103276, https://doi.org/10.1016/j.erss.2023.103276.

[9] Dechezleprêtre, A. et al. (2022), “Searching for carbon leaks in multinational companies”, Journal of Environmental Economics and Management, Vol. 112, p. 102601, https://doi.org/10.1016/j.jeem.2021.102601.

[24] European Union (2020), Regulation (EU) 2020/852 of the European Parliament and of the Council of 18 June 2020 on the establishment of a framework to facilitate sustainable investment, and amending Regulation (EU) 2019/2088.

[10] Fowlie, M. and M. Reguant (2018), “Challenges in the Measurement of Leakage Risk”, AEA Papers and Proceedings, Vol. 108, pp. 124-129, https://doi.org/10.1257/pandp.20181087.

[19] Gatti, L., P. Seele and L. Rademacher (2019), “Grey zone in – greenwash out. A review of greenwashing research and implications for the voluntary-mandatory transition of CSR”, International Journal of Corporate Social Responsibility, Vol. 4/1, https://doi.org/10.1186/s40991-019-0044-9.

[21] Ginder, W., W. Kwon and S. Byun (2019), “Effects of Internal–External Congruence-Based CSR Positioning: An Attribution Theory Approach”, Journal of Business Ethics, Vol. 169/2, pp. 355-369, https://doi.org/10.1007/s10551-019-04282-w.

[4] Gonzalez, M., K. Taddonio and N. Sherman (2015), “The Montreal Protocol: how today’s successes offer a pathway to the future”, Journal of Environmental Studies and Sciences, Vol. 5/2, pp. 122-129, https://doi.org/10.1007/s13412-014-0208-6.

[16] IEA (2024), World Energy Outlook 2024, IEA, Paris, https://www.iea.org/reports/world-energy-outlook-2024.

[6] IEA (2021), Net Zero by 2050 - A Roadmap for the Global Energy Sector, IEA, Paris, https://www.iea.org/reports/net-zero-by-2050.

[20] OECD (2025), “Protecting and empowering consumers in the green transition: Misleading green claims”, OECD Digital Economy Papers, No. 375, OECD Publishing, Paris, https://doi.org/10.1787/12f28e4f-en.

[7] OECD (2025), The Role of Shipbuilding in Maritime Decarbonisation: Impacts of Technology Developments and Policy Measures, OECD Publishing, Paris, https://doi.org/10.1787/0c8362c0-en.

[5] OECD (2024), “OECD Agenda for Transformative Science, Technology and Innovation Policies”, OECD Science, Technology and Industry Policy Papers, No. 164, OECD Publishing, Paris, https://doi.org/10.1787/ba2aaf7b-en.

[17] OECD (2022), Recommendation of the Council on Environmental Information and Reporting, OECD Legal Instruments, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0471.

[26] OECD (2020), Beyond Growth: Towards a New Economic Approach, New Approaches to Economic Challenges, OECD Publishing, Paris, https://doi.org/10.1787/33a25ba3-en.

[18] OECD (2018), “Blue Sky perspectives towards the next generation of data and indicators on science and innovation”, in OECD Science, Technology and Innovation Outlook 2018: Adapting to Technological and Societal Disruption, OECD Publishing, Paris, https://doi.org/10.1787/sti_in_outlook-2018-en.

[13] OECD (2015), Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental Development, The Measurement of Scientific, Technological and Innovation Activities, OECD Publishing, Paris, https://doi.org/10.1787/978926.

[14] OECD (2012), OECD Environmental Outlook to 2050: The Consequences of Inaction, OECD Publishing, Paris, https://doi.org/10.1787/9789264122246-en.

[2] OECD (2011), Towards Green Growth, OECD Green Growth Studies, OECD Publishing, Paris, https://doi.org/10.1787/9789264111318-en.

[15] OECD (forthcoming), Environmental Outlook on the Triple Planetary Crisis: Stakes, Evolution and Policy Linkages.

[25] Stern, N. (2008), “The Economics of Climate Change”, American Economic Review, Vol. 98/2, pp. 1-37, https://doi.org/10.1257/aer.98.2.1.

[8] Stock, J. (2022), Climate Change and the Macroeconomy: Macro Transition Risk and Uncertain Climate Policy, https://scholar.harvard.edu/stock/presentations/climate-change-and-macroeconomy-macro-transition-risk-and-uncertain-climate (accessed on 22 December 2023).

[1] WCED (1987), Our Common Future, World Commission on Environment and Development, (Brundtland Commission), Oxford University Press, Oxford, United Kingdom, https://global.oup.com/academic/product/our-common-future-9780192820808?cc=fr&lang=en&.

[3] Whitesides, G. (2020), Learning from Success: Lessons in Science and Diplomacy from the Montreal Protocol, http://www.ScienceDiplomacy.org.

[23] Yong, S. et al. (2024), “Management Practices and Climate Policy in China”, Journal of the Association of Environmental and Resource Economists, Vol. 11/5, pp. 1065-1100, https://doi.org/10.1086/729013.

Notes

← 1. Examples include the influential Stern review on the economics of climate change (Stern, 2008[25]), which provided economic policy-oriented insights on the implications of alternative courses of action in response to the climate challenge. This study highlighted not only the important of incorporating natural science into economic analysis for effective measurement of costs of environmental change, but also social science and humanistic debates embedded in the economic analysis of factors such as risk, uncertainty and discounting. Science provides tools for understanding the relationships of mutual interdependence across actors in the economy and society and how these can shape incremental change and transformations.

← 2. The historically proven transformational potential of science and innovation has been taken by the most “techno-optimists” as conclusive evidence that scientific and technical progress can be entirely relied upon to offset the loss of natural capital that results from economic development and growth. Those on the more pessimistic end argue that, despite the emergence of short-term solutions, sustained growth is ultimately an unattainable goal, pointing to the fact that the economy, where all factors of production are made of materials and use energy, is a subsystem of the larger finite and nongrowing ecosystem that is the Earth and the Solar System. Indeed, these processes are governed by the laws of thermodynamics, which ensure that all resources are turned back into wastes, in a more “entropic”, or disordered (and therefore often polluting), state (OECD, 2020[26]). “Techno-pessimists” also point to the experience of technologies resulting in unintended environmental effects of some dire consequences. This report does not attempt to enter into such a debate, to which there is probably no satisfactory consensus answer at present, but provides a key driver for current and future scientific and policy enquiry. Regardless of the view one takes on the possibilities of economic growth, the fact that such debate is ongoing is actual proof of the importance for societies to understand science and innovation and its impact, actual and potential, on the natural environment.

← 3. Definitions and taxonomies, such as the EU green taxonomy of sustainable activities (European Union, 2020[24]), are not necessary solely for measurement but also help guide the decisions of investors, consumers, enforcement agencies and others. It is necessary to pay attention to the diverse set of relevant natural resource and ESOs and taxonomies, that are used by several other practitioners. As science and innovation are economic activities, they fall within the scope of application of ESO taxonomies. Companies, universities, research institutes and other stakeholders in the STI space are looking into this and what it implies for them, especially when it comes to action and disclosure to investors, funders, users, consumers, regulators or those authorities who provide the funding. It is therefore relevant for STI measurement to take heed of this opportunity.

2. Science and new technology development for energy and the environment

Abstract

This chapter presents a comprehensive range of indicators mapping a diverse set of channels through which the generation of scientific and technological knowledge on energy and the environment may contribute to sustainable growth. It provides international comparisons, trends, and structural analysis through indicators that depict those channels. In addition to measures of scientific output, collaboration, patents and research and development, this chapter includes newly developed measures of scientific output relevant to sustainable energy and environmental goals, contributions to the Intergovernmental Panel on Climate Change reports, and science underpinning patented innovations on climate change related technologies.

In brief

This chapter seeks answers to questions such as whether the rate of production of scientific knowledge corresponds to the scale and urgency of energy and environmental challenges or what are the characteristics of science that supports environmental and energy policy objectives and the researchers who produce it. Through existing and newly developed indicators, which provide evidence of major trends and structural issues and enable the comparison of countries' scientific and technological capabilities and outcomes, this chapter finds that:

• Several recent trends in global scientific publishing are particularly marked in energy research. OECD countries’ overall scientific output is relatively unspecialised in the area of energy research, and their citation performance is appreciably lower than the world average. In contrast, the People’s Republic of China (hereafter “China”) is by a significant margin the largest contributor to scientific publication output in energy journals. Its scientific output is relatively specialised in this domain, and its rate of highly cited publications is similar to the world average.

• Articles published in specialised scientific energy and environment journals are only a subset of all science relevant to energy and environmental goals. Applying a new AI-based classification tool indicating the societal goal relevance individual scientific publications reveals that the share of scientific publications that contribute to energy and environmental goals is close to 28%.

• When it comes to all energy and environmentally relevant publications (not only those in specialised journals) and focusing on the 10% most highly cited scientific publications worldwide, China accounts for 40% of the world’s scientific output publications (up from 15% in 2012), followed by the European Union and the United States, with 15% and 10%, and down from 27% and 23%, respectively. Across the OECD, with few exceptions, most countries experienced a drop in the share of publications relevant to energy and environmental goals between 2012 and 2022.

• There are few available indicators of science and innovation-related human resources in this area but those that are available offer some initial insights:

• In European countries, enrollment in doctoral education in formal environmental fields is rather low, at close to 1%. However, dissertation work relevant to energy and environmental goals might be significantly higher, at close to 19%, based on detailed data available for France.

• The highest shares of researchers who claim that their work is most relevant to Sustainable Development Goals (SDGs) clustered under the “planet” umbrella1 are found in agricultural and veterinary sciences (49%), natural sciences (28%) and engineering and technology (19%). The highest shares of researchers who consider their work relevant to SDG 7 (Affordable and clean energy) can be found in engineering and technology (16%) and natural sciences (5%).

• At 16%, start-ups in the energy, resources and sustainability sector are the second-most likely to be founded by PhD holders, only after health and biotechnology start-ups (25%), the paradigmatic model of academic entrepreneurship.

• The Intergovernmental Panel on Climate Change (IPCC) Assessment Reports are an example of the key role played by science in building broad societal awareness of environmental challenges, their causes and potential mitigation options. The reports show increasing reliance on science produced across a broader group of countries and a particularly high rate of international scientific collaboration.

• International scientific collaboration is higher for scientific publications related to energy and environmental goals than in other fields. From this higher base, collaboration has been increasing in line with broader trends for all scientific publications.

• Patent data provide an indication of technological development suitable for addressing environmental challenges:

• Between 2010 and 2020, environmental patents filed in at least two intellectual property (IP) offices, one of which is a top 5 IP office, (IP5, an indicator of high-quality patenting) by inventors based in China experienced an extraordinary 604.97% increase, surpassing the United States (15.84% increase) and closely approaching the European Union (12.04%)).

• Nearly 25% of all Patent Cooperation Treaty (PCT) patent filings in China are related to environmental outcomes, compared to only 10% in the United States. The two countries had similar environment-related patent filing intensity a decade ago.

• New inventions that address environmental challenges build on science to a greater extent than those based on existing high-carbon technologies. Although low-carbon technology patents rely only marginally more on non-patent literature, they account for nearly six times as many publications in their citations than high-carbon patents and those with ambiguous carbon-emission impacts. Those rely relatively more on trade literature. Specifically:

• Nearly 40% of scientific publications cited in low-carbon patents are by US-based authors, followed by China with 13% and Germany and Japan with 8% each. This distribution is expected to change, as cited publications by authors based in China are, on average, four years more recent than the average.

• In addition to engineering (16%), the key most-cited fields by low-carbon patents are chemistry (15%) and materials science (12%). As a share of all citations, computer science (7%) is 5 percentage points more important for low-carbon than high-carbon patents.

• The share of venture capital for environment-related start-ups continues to grow in the European Union and China but stagnates across the OECD area.

• Research and development (R&D) in the business sector contributing to energy applications is significantly higher than R&D conducted by energy and other utilities, which account, on average, for less than 1% of the total. In the United States, nearly 5% of business R&D is oriented to energy applications and 2% to environmental protection. The information and communication technology (ICT) and transport equipment industries are significant contributors to energy R&D.

• According to publicly disclosed accounts, companies in the alternative energy and electricity sectors have seen robust growth in R&D expenditure since 2017, second only to the software, computer and electronics sector.

Science and its contribution to the energy and green transition

Measurement rationale

Understanding and responding to complex and global environmental challenges like climate change, pollution or biodiversity loss largely depends on our ability to generate scientific knowledge (World Meteorological Organisation, 2023[1]). Scientific research helps build consensus on the state of planetary ecosystems, its key drivers and the scenarios for future development and human intervention. This knowledge can also inform governments and citizens with a balanced assessment of the potential impact of environmental issues (Rogelj, 2023[2]). Scientific research also plays a critical role in pushing the boundaries of knowledge that applied researchers, engineers, and designers can draw upon to develop new, viable technological solutions (Perrons, Jaffe and Le, 2020[3]). By enhancing our understanding of fundamental principles and charting unexplored pathways, basic science provides the means to develop new approaches for tackling otherwise intractable technical and socio-economic problems using current technology. Without such advances, social and political willingness to embrace environmental objectives and act on them would be severely diminished.

As discussed in Chapter 1, measuring science and research relevant to environmental sustainability and effective use of natural resources is not a trivial endeavour. Demonstrating relevance can be more straightforward for applied research, which is oriented towards specific, practical goals. However, it takes considerable effort to trace how fundamental research into principles and facts of nature and society can be relevant for future technologies and policy decisions. This chapter explores multiple channels through which these knowledge generating activities operate and provides a range of key indicators, drawing on a diverse range of data sources.

Indicators of scientific activity relating to energy and the environment

A first step in assessing the relevance of scientific output to societal goals related to energy and the environment is to focus on the academic journals that specialise in those areas. As detailed in Box 2.1, the fields of agricultural and biological sciences; environmental science; earth and planetary science; and energy within the All Science Journal Classification (ASJC) are most explicitly connected with energy and environmental sustainability. Bibliometric analysis shows that in 2023 China was the largest country contributor to scientific publishing in these areas, surpassing the European Union and fast approaching the combined value of OECD countries, followed by the United States, India and Russia (Figure 2.1).

Box 2.1. Bibliometrics and scientific journals in energy and environment fields

The scientific peer review system and the body of scholarly publications it generates help provide the basis on which bibliometrics can be applied to the study of scientific research. Quantitative studies on research publications can draw on the contents and information contained in publications – which meet defined review and publication criteria – to analyse, under some assumptions, multiple dimensions of scientific production and dissemination. Specifically:

• Indexed information on scientific documents helps investigate sources of scientific knowledge through the identity and affiliation of authors and the references contained in a document.

• Scientific collaboration can be potentially gauged by the extent of co-authorship and/or the engagement of multiple institutions.

• The relevance of the research to the broader scientific community may be inferred, in part, from the extent a publication is cited by other documents or the visibility of the title, e.g. the journal, in which a given document is published, based on its past citation record.

The interpretation of bibliometric analysis is contingent on a series of norms and incentives that vary across sectors and knowledge domains. It can also evolve over time. For example, not all scientific discoveries and research results are published in a well-defined list of international scientific journals where they can be read and cited by other researchers and a possibly wider user community. Like other forms of administrative data sources, bibliometric data do not exist to serve statistical purposes. While this does not invalidate their relevance for the statistical analysis of science, it is important to remember this key feature and apply a degree of caution when designing and interpreting bibliometric indicators.

Inferring relevance from journal classification

The ASJC system is used in the Scopus database, a bibliographic data source used by the OECD, to classify journals and conference proceedings under broad subject areas that are further divided into groups and subfields. The ASJC fields of agricultural and biological sciences; environmental science; earth and planetary science; and energy are explicitly connected with energy and environmental sustainability. Publications in journals classified in these fields can be presumed potentially relevant to sustainable growth. This approach, however, misses out on relevant scientific publications in other journals, not only those of a multidisciplinary nature. The contribution of basic science, for instance, would not necessarily be captured using this approach, thus calling for other approaches presented elsewhere in this publication.

Source: Authors, partly based on OECD and SCImago Research Group (CSIC) (2016[4]), Compendium of Bibliometric Science Indicators, https://web-archive.oecd.org/2016-09-29/415063-Bibliometrics-Compendium.pdf.

Figure 2.1. Scientific publication volume in energy and environment-themed journals, 2023

Main contributing economies to journals in agricultural and biological sciences, environmental, earth and planetary sciences and energy fields

Note: Fractional counts of publications in journals in the All Science Journal Classification (ASJC) fields of agricultural and biological sciences; environmental science; earth and planetary science; and energy. Publications have been fractionally attributed to geographical areas and scientific domains based on the authors' institutional affiliations and Elsevier’s ASJC tagging of journals for citeable documents.

Source: OECD (n.d.[5]), Science and bibliometric indicators, https://www.oecd.org/en/data/datasets/science-and-bibliometric-indicators.html; OECD calculations based on Scopus Custom Data, Elsevier, March 2023, accessed from OECD STI.Scoreboard, https://stip.oecd.org/stats/SB-StatTrends.html?i=FPUBS_21_NBFRAC,FPUBS_NBFRAC,FPUBS_23_NBFRAC,FPUBS_11_NBFRAC,FPUBS_19_NBFRAC&v=8&t=2021&r=3.

Statlink

As a share of total domestic publication output (Figure 2.2), the rate of scientific publishing in energy and environmental fields is highest in Costa Rica (34%) and Argentina (26%). This reflects, in particular, the relative importance of agriculture for these economies and the extent to which their scientific production is oriented towards meeting economic and social needs.

Figure 2.2. Scientific publication intensity in energy and environment-themed journals, 2023

Economies with the largest shares of scientific publications in agricultural and biological sciences, environmental, earth sciences and energy journals

Note: Domestic shares of fractional counts of citeable publications in journals in the ASJC fields of agricultural and biological sciences; environmental science; earth and planetary science; and energy, relative to all domestic publications. Publications have been fractionally attributed to geographical areas and scientific domains based on the authors' institutional affiliations and Elsevier’s ASJC tagging of journals.

Source: OECD (n.d.[5]), Science and bibliometric indicators, https://www.oecd.org/en/data/datasets/science-and-bibliometric-indicators.html. OECD calculations based on Scopus Custom Data, Elsevier, March 2023, accessed from OECD STI.Scoreboard, https://stip.oecd.org/stats/SB-StatTrends.html?i=TOP10FPUBS_19_NBFRAC,TOP10FPUBS_23_NBFRAC,TOP10FPUBS_21_NBFRAC&v=8&t=2021&r=4 (link includes data for a broader set of economies).

Statlink

Examining trends in environmental and energy-related journal fields over the 2009-23 period among selected economies (Figure 2.3) shows that scientific publishing is only consistently on the rise as a proportion of total publishing in the area of environmental science, in what seems to reflect a growing preoccupation with the implications of natural resource utilisation, pollution and climate change. The share of agricultural and biological sciences has declined in the European Union, the United States and across OECD countries, while it has seen robust growth in China. Among those presented, China is the only major economy where the relative importance of all these fields, including energy, has increased over the period under examination.

China is by a significant margin the largest contributor to scientific publication output in energy journals, a field in which it is relatively specialised and has a rate of highly cited publications approximately similar to the world average. In contrast, the EU’s overall scientific output is relatively unspecialised in energy, and its citation performance is appreciably below the world average (Figure 2.4). European Commission’s analysis of research specialisation in each of the ‘Horizon 2020 Societal Grand Challenges’ finds that, overall, the EU is more specialised in publications related to health and less specialised in publications on secure societies and energy. In terms of the climate action, environment, resource efficiency and raw materials challenge, EU’s specialisation is around the global average, and remains relatively unchanged since 2000 (European Commission, 2024[6]).

Figure 2.3. Scientific publishing trends in the fields of energy and environment, selected economies, 2009-23

Share of domestic scientific publications in agricultural and biological sciences, environmental, earth and planetary sciences and energy journals

Note: Shares of each economy’s publications in journals in ASJC fields of agricultural and biological sciences; environmental science; earth and planetary science; and energy. Publications have been fractionally attributed to geographic territories and scientific domains based on authors' institutional affiliations and Elsevier’s tagging of journals for citeable documents.

Source: OECD (n.d.[5]), Science and bibliometric indicators, https://www.oecd.org/en/data/datasets/science-and-bibliometric-indicators.html. OECD calculations based on Scopus Custom Data, Elsevier, March 2023

Statlink

Figure 2.4. Specialisation and citation impact in scientific publications in the energy field, 2023

Total publications displayed as bubble size

Note: "Top-cited publications" are the 10% most-cited papers normalised by scientific field and type of document (articles, reviews and conference proceedings). The Scimago Journal Rank indicator is used to rank documents with identical numbers of citations within each class. This measure is a proxy indicator of research excellence. Estimates are based on fractional counts of documents by authors affiliated to institutions in each economy. Documents published in multidisciplinary/generic journals are allocated on a fractional basis to the ASJC codes of citing and cited papers. The relative specialisation indicator has been calculated as the ratio of a given field's share in a country's total scientific production, relative to the world's equivalent. A ratio higher than 1 signifies a high degree of specialisation, with the field's share in that country exceeding the relative importance of the field in overall global scientific output, as captured by the Scopus database. Figures have been rounded. Instances with too few documents in a given economy and field have been suppressed.

Source: OECD calculations based on Scopus Custom Data, Elsevier's Scopus Custom Data, Version 1.2025; and Scimago Journal Rankings.

Statlink

Self-reported measures of the relevance of science to energy and environmental goals

Given the limitations of determining which publications are relevant to environmental sustainability based on ASJC fields, there is a need for alternative approaches. One alternative is to ask the researchers themselves about the societal goals with which their research aligns the most. The OECD International Survey of Science (ISSA) conducted in 2021 collected data from 3 091 scientific researchers, who were asked a range of questions on their working conditions, society engagement, the impact of the coronavirus (COVID‑19) pandemic on their work, career prospects, and the alignment of their research with the United Nations (UN) SDGs. Some 94% of respondents (2 911 individuals) answered the specific question concerning the relevance of their research to one or more SDGs. Only 8% reported that their work had no relevance to the SDGs. While not necessarily representative of the entire research workforce, the survey reveals new insights regarding the orientation of science fields to various societal goals, as defined by the Sustainable Development Agenda, as well as about various characteristics of the researchers who engage in research supporting environmental sustainability and energy (Figure 2.5).

Figure 2.5. Distribution of self-reported most relevant SDG to their research among ISSA 2021 respondents

Distribution of SDGs reported as most relevant, clustered by group, as percentage of respondents in each field

Note: Results are based on ISSA 2021 survey participants’ responses to the question: “Which SDG is your scientific or research activity most relevant for?” Planet = SDG 6: Clean water and sanitation, SDG 12: Responsible consumption and production, SDG 13: Climate action, SDG 14: Life below water, SDG 15: Life on land. The prosperity cluster has been separated into prosperity – energy = SDG 7 (Affordable and clean energy) and prosperity – other = SDG 8 (Decent work and economic growth), SDG 9 (Industry, innovation and infrastructure), SDG 10 (Reduced inequalities), SDG 11 (Sustainable cities and communities). Analysis of responses indicating “quality education” as objective reveals that these do not necessarily correspond to research on education but to scholarship-driven research. This goal is more commonly reported by researchers in higher education with teaching duties, who are involved in basic research and especially in the fields of arts and humanities and social sciences. Unweighted results based on 2 908 responses. Approximately 8% of respondents indicate “no SDG relevance”.

Source: OECD (2021[7]), OECD International Survey of Science, http://oe.cd/issa.

Statlink

As might be expected, the highest shares of researchers who claim that their work is most relevant to the SDGs clustered under the “planet” umbrella are found in agricultural and veterinary sciences (49%), natural sciences (28%) and engineering and technology (19%). The highest shares of researchers who consider their work relevant to SDG 7 (Affordable and clean energy) can be found in engineering and technology (16%) and natural sciences (5%). According to the survey, other fields of science only have a negligible energy share.

Basic research has long been recognised as the “pacemaker of technological progress” (Bush, 1945[8]). By its very definition, the eventual application of basic research is often difficult to predict, but it is known to play a fundamental role in “clean tech” patented technologies, especially in the “deep tech” subset (Dalla Fontana and Nanda, 2023[9]). For example, advances in solid-state physics and semiconductor research in the mid-20th century paved the way for the discovery of the photovoltaic effect in silicon, which underpins today’s solar photovoltaic industry (Chapin, Fuller and Pearson, 1954[10]; Green, 2000[11]). Similarly, the development of atmospheric physics and early computational models in the 1960s laid the groundwork for modern climate modelling, which has become essential for understanding anthropogenic climate change and guiding mitigation and adaptation strategies (IPCC, 2023[12]).

These examples illustrate how knowledge generated without immediate commercial or environmental objectives can yield transformative spillovers, accelerating the shift towards low-carbon economies. According to the OECD International Survey of Science, only 21% of the researchers whose work is relevant to the “planet” SDG category report that it is purely basic research. However, the combined shares of basic research and basic and applied research are some of the highest in both the “planet” and the “prosperity – energy” categories, suggesting that basic research, including in combination with applied research, plays an important role in advancing these particular societal goals (Figure 2.6).

Figure 2.6. Orientation of research according to SDG goal relevance

Distribution of modalities of research orientation according to the most relevant SDG goal

Note: The information on research modalities and scientific activity is based on multiple question items. The indicator is based on responses to items on basic and applied research, as defined in the OECD Frascati Manual. The category of “neither basic nor applied research” may include experimental development and related activities, as well as teaching. Planet = SDG 6: Clean water and sanitation, SDG 12: Responsible consumption and production, SDG 13: Climate action, SDG 14: Life below water, SDG 15: Life on land. The prosperity cluster has been separated into prosperity – energy = SDG 7 (Affordable and clean energy) and prosperity – other = SDG 8 (Decent work and economic growth), SDG 9 (Industry, innovation and infrastructure), SDG 10 (Reduced inequalities), SDG 11 (Sustainable cities and communities).

Source: OECD (2021[7]), OECD International Survey of Science, http://oe.cd/issa.

Statlink

New indicators of energy and “green” science

Available classification methods have several limitations, which constraints their use as a basis for indicators relevant to societal goals. While the use of “key terms” that relate to those goals to query the titles and abstracts of scientific publications may offer some useful insights into relevance of scientific publications to certain topics (e.g. Box 2.2), this approach often understates the contribution of basic science to attaining those goals.

Given these limitations, the OECD has developed an SDG classification methodology that makes it possible to analyse energy and environmental relevance of the full body of scientific publications regardless of the type of journals documents are published in and accounting for the subtle distinction between what research is about and its potential relevance and applicability. As summarised in Box 2.3, the new methodology brings together the potential of surveys, bibliographic data, and artificial intelligence (AI) tools for analysing descriptions of research work.

Box 2.2. Measurement case study: Analysis of space-based “green” research by the Danish Ministry of Higher Education and Science

Policymakers are often interested in how two different domains interact and assessing how competencies in one particular scientific domain contribute to advances in environmental science. The Danish Ministry of Higher Education and Science (UFM) conducted a bibliometric analysis of Denmark's space-based “green” research (Ministry of Higher Education and Science, 2022[13]), assessing its role in the country's broader green transition efforts. The report analysed the volume of relevant scientific publications, distribution across disciplines, specialisation, collaboration patterns and impact.

The study identified Denmark’s space-based green research by combining two search strategies implemented on a commercial database to identify research output that integrates space technology with environmental science. The “green research” identification follows the Ministry’s framework, covering sustainability, climate change, biodiversity, circular economy and renewable energy. The “space-based” query uses relevant keywords such as "satellite," "Earth observation," "remote sensing," and many others. The search used information on titles, abstracts and author keywords. Both the space keywords and the publications retrieved by the overall search strategy were subject to quality assurance by eight Danish universities and the Danish Meteorological Institute.

Source: Authors, based on Ministry of Education and Research (2022[13]), Rumbaseret grøn forskning – Bibliometrisk analyse af Danmarks rumbaserede grønne forskning September 2022, www.ufm.dk.

Applying the SDG classification reveals that the share of scientific publications that contributes to energy and environmental goals remained close to approximately 28% globally between 2008 and 2018 but is not uniquely concentrated in a few fields (Figure 2.7). The SDG-based classification makes it possible to revisit the classification of scientific publications by ASJC field and to examine the degree to which different fields are relevant to natural resource management and environmental sustainability. The shares of relevant publications are indeed the highest within the initially selected ASJC fields, namely energy, environmental sciences, agricultural and biological sciences and earth and planetary sciences. However, the results show that considerable scientific output within other fields was being ignored, particularly in chemical engineering and chemistry journals, which display a share of energy- and environment-relevant output of over 40%. Contributions of less than 10% are only found in the different health and medical science fields.

Box 2.3. Introducing the new OECD SDG research relevance classifier

Understanding the relevance of research to a societal goal is an important but rather different task than examining the subject or field of research. The text contained in scientific publication abstracts and R&D project descriptions lends itself to analysing subject matter in ways that significantly improve analysis that relies on journal classifications. However, it does not necessarily provide a good indication of what research is relevant for. Machine learning classifiers typically rely on topic similarity of text-based descriptions of the work but often lack explicit input into the classification process from those in a position to assess the relevance to goals such as the SDGs. This can result in understating the relevance of basic research or missing out on how goals can depend on each other.

In response to this challenge, the OECD has developed an SDG classifier using information available in the responses to the ISSA survey conducted in 2021, whose results are presented earlier in this chapter. Self-assessments of relevance have been used to train a classifier algorithm. A subset of the respondents, 69% (2 014 individuals), were successfully matched to their author profiles in Scopus 2024. A total of 11 719 publications from 2000-21 were retrieved from Scopus and associated with these authors. Each publication was labelled with its relevant SDG based on the individual respondent's ISSA survey response.

The SDG classifier relies on the SciBERT large language model, which has been trained on scientific texts, making it more adept at understanding specialised terminology. It allows for multi-label classification, reflecting the fact that the SDGs are not mutually exclusive. In multi-label classification, each label's probability is independent of the other. Since the survey respondents were able to select the most relevant SDG, including the possibility of selecting none, normalisation is performed to ensure the distribution of probabilities adds up to 100% across all possible outcomes. The thus trained classifier can be used for the SDG classification of a wide variety of texts, including scientific publications, thesis abstracts and R&D funding awards from the OECD Fundstat database.

This publication focuses on a selected group of SDGs that refer to energy (SDG 7: Clean and affordable energy) and the environmental sustainability “planet” cluster (SDG 6: Clean water and sanitation; SDG 12: Responsible consumption and production; SDG 13: Climate action; SDG 14: Life below water; SDG 15: Life on land) (hereafter referred to as “environmental”) goals. The classifier assigns probabilities to each object (publications and PhD theses in this chapter, R&D awards in Chapter 4) and uses them for fractional counting and tagging. A publication is thus tagged as relevant to energy and environmental goals if the predicted SDG cluster probability is highest for the combination of SDG 7 (Affordable and clean energy) and the SDG “planet” cluster. These are described as “energy and environmental” SDG-related publications for brevity.

Source: Authors, based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods”.

Figure 2.7. Scientific publications relevant to energy and environmental goals, by field of journal, 2022

As a percentage of all publications within each journal ASJC field domain

Note: A publication is tagged as relevant to energy and environmental goals if the predicted SDG cluster probability is highest for the combination of SDG 7: Affordable and clean energy and the “planet” cluster (SDG 6: Clean water and sanitation; SDG 12: Responsible consumption and production; SDG 13: Climate action; SDG 14: Life below water; SDG 15: Life on land).

Source: OECD calculations based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods” and on Scopus Custom Data, Elsevier, Version 1.2024.

Statlink

Figure 2.8. Distribution of global scientific publications by relevance to societal goals, 2022

Note: A publication is tagged as relevant to energy and environmental goals if it has the highest aggregated probability for SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes.

Source: OECD calculations based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods” and Scopus Custom Data, Elsevier, Version January.2024.

Statlink

Figure 2.8 shows the estimated distribution of world’s total scientific publication output for 2022 across the different SDGs, including the “no SDG relevance” case. The energy and environment SDGs account for approximately 28% of scientific production. Among those, SDG 7 (Affordable and clean energy) is the most contributed SDG at nearly 7% of the total, followed by SDG 13 (Climate action). The results indicate that, contrary to previous work, the vast majority of scientific publication output (94%) is relevant to at least one SDG and, with exceptions such as SDG 3 (Good health and well-being) and SDG 9 (Industry, innovation and infrastructure), science efforts are broadly spread across SDGs, which in many cases overlap with each other.

The number of scientific publications deemed most likely to contribute to the energy and environmental sustainability transition increased by almost 100% from 2008 to 2022, from 500 000 to almost 1 million, thereby growing slightly over the rate of all scientific publications (Figure 2.9). However, through the COVID-19 crisis period, the share of publications relevant to energy and environmental goals declined moderately

Figure 2.9. Trend in scientific publications relevant to energy and environmental goals, 2008-22

Thousands of publications, allocated on a fractional count basis

Note: Thousands of publications, allocated on a fractional count basis. A random sample comprising 10% of the total publications was tagged for relevance to environmental sustainability and energy. The results were then extrapolated to represent the entire population of publications.

A publication is tagged as relevant to energy and environmental goals if it has the highest aggregated probability for SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes.

Source: OECD calculations based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods” and Scopus Custom Data, Elsevier, Version 1.2024.

Statlink

There have been major changes in the contribution of the largest global economies to energy and environment-relevant output. The United States and the European Union have seen a large decline in the share of relevant publications. The share in China has increased rapidly, and India has also seen a steady increase (Figure 2.10). This implies a reduction in the overall relative contribution of OECD economies to scientific output in this area, over and above the general scientific publication shift that has been taking place.

Within the OECD, most countries have experienced a drop in the share of publications relevant to the energy and environmental goals between 2012 and 2022, except Iceland and Türkiye (Figure 2.11. ). Latvia, Costa Rica, and Estonia exhibit some of the largest shares, with values of well over 30%. The United States, the Netherlands and Israel occupy the lowest ranks within the OECD, with a share of relevant publications between 14% and 18%. The EU average ranks higher than the OECD average.

Figure 2.10. Main contributors to scientific publications relevant to energy and environmental goals, 2008-22

As a percentage of world's total publications relevant to energy and environmental goals

Note: A publication is tagged as relevant to energy and environmental goals if the predicted SDG cluster probability is highest for the combination of SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes.

Source: OECD calculations based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods” and Scopus Custom Data, Elsevier, Version 1.2024.

Statlink

Figure 2.11. Scientific publications relevant to energy and environmental goals, selected economies, 2012 and 2022

OECD countries and the European Union, as a percentage of total publications

Note: A publication is tagged as relevant to energy and environmental goals if the predicted SDG cluster probability is highest for the combination of SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes. See stat.link for more countries.

Source: OECD calculations based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods” and Scopus Custom Data, Elsevier, Version 1.2024.

Statlink

Distinguishing between publications oriented to the energy SDG 7 and other environmental sustainability SDGs under the planet cluster, it is evident that energy is particularly important in the case of Mexico, Japan and Korea for OECD countries, and several oil- and gas-producing countries among selected non-member economies, such as Kazakhstan, the Russian Federation (hereafter “Russia”) and Saudi Arabia (Figure 2.12). There is no case where the volume of scientific publications relevant to the energy SDG surpasses the ensemble of the other environmental SDG goals.

Figure 2.12. Contributions to scientific publications relevant to energy and environmental goals, selected economies, 2022

As a percentage of total publications, OECD and selected economies

Note: A publication is tagged as relevant to energy and environmental goals if it has the highest aggregated probability for SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes. For texts where the probability of the aggregate environment and energy category is the maximum, planet-focused vs energy-focused are further distinguished according to the next highest category probability.

Source: OECD calculations based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods” and Scopus Custom Data, Elsevier, Version 1.2024.

Statlink

Focusing on the top 10% most-cited publications within their fields, China accounts for an even larger share of the world’s total than for all publications (close to 40% in 2022), followed by the European Union and the United States, with 15% and 10% respectively (Figure 2.13). This indicator further demonstrates that China does not only lead in environment-related product manufacturing and exports, but increasingly also in the creation of relevant knowledge (IFC, 2025[15]). Both China and India have substantially increased their share of relevant publications between 2012 and 2022, defying the otherwise almost universal decline among the group of countries that contribute the most to the world's top 10% cited publications. France, the United Kingdom, the United States and Germany experienced the largest declines. International collaboration has become one of the most significant features of scientific discovery and technological development in the 21st century and is a key vector of information exchange (OECD, 2017[16]). International scientific collaboration is particularly important in scientific research relevant to the energy and environmental goals compared to all other goals. This collaboration “premium” for environmental and energy-relevant research has remained significant despite the overall increase in collaboration (Figure 2.14).

Figure 2.13. Main contributors to the top 10% most cited scientific publications relevant to energy and environmental goals, 2012 and 2022

As a percentage of the world's top 10% most-cited energy and environment-relevant scientific publications

Note: A publication is tagged as relevant to energy and environmental goals if it has the highest aggregated probability for SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes.

Source: OECD calculations based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods” and Scopus Custom Data, Elsevier, Version 1.2024.

Statlink

Figure 2.14. Trends in international scientific collaboration relevant to energy and environmental goals, 2000-2022

As a percentage of all publications

Note: International collaboration refers to publications co-authored among institutions in different countries. Estimates are computed for each country by counting documents for which the set of listed affiliations includes at least one address within the country and one outside. Single-authored documents with multiple affiliations in different countries count as institutional international collaboration. A publication is tagged as relevant to energy environmental goals if it has the highest aggregated probability for SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes.

Source: OECD calculations based on Scopus Custom Data, Elsevier, Version 1.2024, April 2024.

Statlink

The pattern holds across most of the surveyed countries. However, in some, the difference in collaboration rates between research relevant to environmental sustainability and energy and other research is small (Figure 2.15). As Figure 2.16 shows, international collaboration within the scientific domain relevant to environmental sustainability has increased over time in all countries but India and the Russian Federation.

Figure 2.15. International scientific collaboration relevant to energy and environmental goals, selected economies, 2022

As a percentage of domestically authored documents, fractional counts, selected economies

Note: International collaboration refers to publications co-authored among institutions in different countries. Estimates are computed for each country by counting documents for which the set of listed affiliations includes at least one address within the country and one outside. Single-authored documents with multiple affiliations in different countries count as institutional international collaboration. A publication is tagged as relevant to energy and environmental goals if it has the highest aggregated probability for SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes.

Source: OECD calculations based on Scopus Custom Data, Elsevier, Version January 2024.

Statlink

Figure 2.16. Intensity of international collaboration in energy and environment relevant scientific publications, selected economies, 2012 and 2022

As a percentage of domestically authored documents, fractional counts

Note: International collaboration refers to publications co-authored among institutions in different countries. Estimates are computed for each country by counting documents for which the set of listed affiliations includes at least one address within the country and one outside. Single-authored documents with multiple affiliations in different countries count as institutional international collaboration. A publication is tagged as relevant to energy and environmental goals if it has the highest aggregated probability for SDG 7: Affordable and clean energy and the SDGs under the “planet” cluster. See Box 2.3 and chapter notes.

Source: OECD calculations based on Scopus Custom Data, Elsevier, Version January 2024.

Statlink

Research and innovation workforce capabilities

Measurement rationale

Progress towards achieving sustainable growth is entirely dependent on the level of relevant knowledge and skills within a population. The availability and productive engagement of specialised talent underpins all possible contributions of science and innovation, as laid out in Chapter 1. There is a fast-growing literature on skills required for the environmental sustainability transition and measuring their current and expected labour market footprint. For example, across OECD countries, 20% of the workforce is estimated to be employed in “green-driven” occupations (Figure 2.17), a category that comprises not only jobs that directly contribute to emission reductions but also those that are not currently but are likely to be, in demand because they provide goods and services needed for environment-related activities (OECD, 2024[17]). Some 14% of “green-driven” employment as defined in the OECD Employment Outlook 2024 is deemed to be in new and emerging occupations that did not previously exist and, therefore, imply a degree of innovation.

Figure 2.17. Green-driven and greenhouse-gas-intensive occupations, selected economies, 2015-19

As a percentage of all employment, period average

Note: Data refer to the average for 2015-19, except for Canada: 2017-19 and New Zealand: 2018. OECD: unweighted average.

Source: OECD estimates based on Version 24.1 of the O*NET Database and the following country-specific sources: Australian Labour Force Survey; Canadian Labour Force Survey; Japanese Labour Force Survey; New Zealand: Household Labour Force Survey; United States: Current Population Survey; All other countries: EU Labour Force Survey.

Statlink

The OECD Employment Outlook 2024 ([18]) notes that transitioning from greenhouse gas (GHG)-intensive to innovative, “green-driven” occupations may be significantly more challenging for low-skilled workers. Ensuring sufficient and appropriate training for low-skilled workers will be paramount to addressing both skill shortages in emerging environment-related industries and supporting the learning needs of low-skilled workers. OECD analysis suggests that most GHG-intensive occupations share similar skill requirements with non-GHG-intensive occupations, suggesting that transitions within highly polluting sectors are feasible with well-targeted reskilling.

However, less attention has been paid to monitoring the nurturing and deployment of advanced research and innovation (R&I) talent that can push scientific and technological boundaries in a direction consistent with sustainable growth objectives. This section explores a limited set of available and experimental indicators that aim to overcome the dual challenge associated with assessing the size of the R&I talent pool, a significant undertaking in its own right (Box 2.4), and obtaining evidence on its capability to contribute to environmental sustainability goals through STI activities.

Box 2.4. Concepts for measuring R&I talent and its contribution to environmental goals

The broad concept of “R&I talent” can be approached as comprising the collectives that are effectively involved or have the potential to contribute to R&D or other related STI activities. Definitions for R&D and innovation are provided in the OECD Frascati and Oslo manuals, respectively. A broad view of R&I talent includes not only those actively involved in R&D (R&D personnel) and innovation but also individuals who, while not actively involved in such activities at any given point, have the potential to do so given their competences, such as educational qualifications, skills and experience.

With international growth in education rates, educational attainment has become an important indicator of the capacity to engage in R&I activities at any level. Although formal qualifications are neither necessary nor sufficient conditions of capability to engage in R&I activities, employers place increasing emphasis on educational accreditation standards. For example, doctoral degree holders have attained the highest level of formal qualification (International Standard Classification of Education [ISCED] 8). They have been specifically trained to conduct original research, and, as a result, as they graduate, they are uniquely well-qualified to contribute to generating scientific knowledge. It is, therefore, relevant to ask whether it is also possible to assess the contribution of these individuals to specific societal goals.

At PhD or other levels, information on the field of education for programmes, as categorised by the ISCED-F classification of fields of education (UNESCO, 2015[19]), provides an additional monitoring dimension for understanding the specific expertise and disciplinary backgrounds within the R&I workforce. Six ISCED-F classification classes are explicitly aligned with the natural environment: 1) environment; 2) environment (not defined); 3) environmental sciences; 4) natural environment and wildlife; 5) environment n.e.c. (not elsewhere classified); and 6) environmental protection technology. However, these cover a relatively minor share of all knowledge domains with proven possible relevance to energy and environmental goals.

Occupational and competence-based frameworks provide a complementary perspective, as both likely “research” and “green”-related jobs have been mapped out in separate thematic exercises. The key challenge ultimately stems from what information can be reliably retrieved from existing and yet-to-be-developed data sources. If data are only generated or made available at highly aggregated occupational levels, it might be misleading to assume that these occupations, across different countries, will have similar propensities to be both environment and R&I-related.

The Research and Innovation Careers Observatory (ReICO) aims to serve as a comprehensive access point for international statistics, analytical tools and resources focused on R&I careers. In collaboration with national governments and various stakeholders, ReICO is developing new methodologies, indicators and insights to support such goals. Working to identify the relevance of R&I talent to specific goals, such as the energy and environmental sustainability transition, will be a likely area of policy user interest in this new monitoring tool.

Source: OECD (n.d.[20]), Research and Innovation Careers Observatory, https://www.oecd.org/en/networks/research-and-innovation-careers-observatory.html.

Indicators of R&I talent development in energy- and environment-related areas

Students enrolled in environment-related doctoral programmes

Statistics on enrolment and graduation by education programmes can be used to map potential additions to the scientific research workforce relevant to sustainability, using the more explicitly connected categories in the ISCED-F classification set out in Box 2.4. Although data with this level of detail are not available for the ensemble of OECD countries, Eurostat data show that within the European Union, Spain, Italy and France had the most students enrolled in PhD-equivalent environment programmes in 2023. As a share of total doctoral students, environmental subjects account for slightly over 1% among EU and associated countries, with Switzerland and Estonia in the lead (Figure 2.19).

Figure 2.18. Students enrolled in doctoral education in environmental fields in EU countries, 2023

Note: Environmental fields comprise those enrolled in the following six Level 4 ISCED-F classes: environment; environment not further defined; environmental sciences; natural environments and wildlife; environment not elsewhere classified; and environmental protection technology. European Union (EU), European Economic Area (EEA) and Switzerland. Only those countries that report across all fields are displayed.

Source: Eurostat (n.d.[21]), “Students enrolled in tertiary education by education level, programme orientation, sex and field of education,” https://ec.europa.eu/eurostat/databrowser/view/educ_uoe_enrt03__custom_15308082/default/table?lang=en&page=time:2021, accessed on 11 May 2025.

Statlink

Figure 2.19. Share of students in doctoral education in environmental fields in EU countries, 2023

As a percentage of the total number of students enrolled across all doctoral education programmes

Note: Data available for EU countries, EEA countries and Switzerland. See notes to the previous figure.

Source: Eurostat (n.d.[21]), “Students enrolled in tertiary education by education level, programme orientation, sex and field of education,” https://ec.europa.eu/eurostat/databrowser/view/educ_uoe_enrt03__custom_15308082/default/table?lang=en&page=time:2021, accessed on 11 May 2025.

Statlink

Experimental indicators of doctoral dissertations related to energy and environmental goals

While classifications of educational programmes provide a similar guide to measurement as journal fields do for classifying scientific output, the newly developed OECD SDG classifier presented earlier (Box 2.3) can be deployed in datasets comprising doctoral dissertation titles, abstracts and full text to determine which ones are relevant to the energy and environment-related SDGs. The application of this SDG classifier is pertinent as it operates with a similar type of research content. Given limited data availability at the international level, this section demonstrates this measurement approach using publicly available data on French doctoral dissertations submitted between 2018 and 2023. Principal component analysis (PCA), a dimensionality-reduction statistical technique with applications in exploratory data analysis and visualisation, helps map the SDG relevance space for doctoral theses in France (Figure 2.20).

Figure 2.20. Patterns of SDG relevance in French doctoral dissertations, 2018-23

Principal component analysis (PCA) biplot of estimated SDG relevance probabilities

Note: The database on which SDG relevance has been estimated comprises information on French doctoral theses, excluding those submitted for purposes of accreditation to become principal investigator, as well as pharmacy, medical doctorate, dentistry or doctoral veterinary dissertations.

Source: OECD analysis based on Government of France (2024[22]), Thèses soutenues en France depuis 1985, https://www.data.gouv.fr/fr/datasets/theses-soutenues-en-france-depuis-1985/.

Statlink

Across the population of PhD candidates, “planet” SDG-related dissertations account for 19%, while the energy share is 5%. The combined 24% value is close to the 22% share found for scientific publications for France. Mapped onto a two-dimensional space, the position of the variables depicting theses’ probabilities for 17 SDGs and no SDG relevance shows clusters that resemble common groupings for SDGs previously introduced but also highlight how the “zero hunger” goal is very closely related to “planet” objectives, in particular, the “Life on land” objective, showing the close connection between science environmental goals and impacts on food systems.

Measures of R&I workforce talent relevant to the energy and green transition

Researchers’ demographic characteristics

The ISSA survey discussed in the previous section also provides insights into the demographic characteristics and working conditions of scientists and researchers whose work is relevant to different SDGs. Understanding patterns can help identify potential trends of under- or over-representation of different social groups and different incentives to pursue research careers with a range of possible sustainability impacts.

Based on the non-probabilistic sample obtained from the survey (Figure 2.21), the age of researchers whose work is most relevant to the environment SDGs is found to be only marginally lower than the average, while scientists whose work is considered most relevant to SDG 7 (Clean and affordable energy) are among the youngest, almost five years below the sample average of 45.9. Younger researchers appear to be more likely to report SDG relevance other than “Quality education” SDGs; this includes respondents who engage in scholarship-driven research rather than research that is necessarily focused on education.

Figure 2.21. Demographic features of ISSA respondents according to SDG relevance

Note: Results are based on ISSA2021 survey participants’ responses to the question: “Which SDG is your scientific or research activity most relevant for?” Planet = SDG 6: Clean water and sanitation, SDG 12: Responsible consumption and production, SDG 13: Climate action, SDG 14: Life below water, SDG 15: Life on land. The prosperity cluster has been separated into prosperity – energy = SDG 7 (Affordable and clean energy) and prosperity – other = SDG 8 (Decent work and economic growth), SDG 9 (Industry, innovation and infrastructure), SDG 10 (Reduced inequalities), SDG 11 (Sustainable cities and communities). Unweighted results based on 2 908 responses.

Source: OECD (2021[7]), OECD International Survey of Science, http://oe.cd/issa.

Statlink

There are no appreciable differences between men and women in terms of the relevance of their work to the environmental SDGs, while men appear to be twice as likely to pursue scientific research work relative to SDG 7 (Affordable and clean energy). Women are particularly more likely to contribute their scientific work to the “Health and society” SDG cluster, while men are more oriented toward the “Economic prosperity” SDG cluster.

Box 2.5. Measurement case study: beyond the R&D workforce – the UK R&I Workforce Survey

The 2022 UK R&I Workforce Survey covers the full diversity of occupations in the R&I workforce, including trainee and experienced researchers, technicians, engineers, R&I leaders and managers. All types of specialisms and sectors (private, public or non-profit) are surveyed. Among other things, the first wave of the workforce survey issued in 2022 supported the business case for the Green Future Fellowship, a programme supporting scientists, researchers and innovators to develop and scale up their breakthrough climate solutions.

While the survey’s definitions do not necessarily align with the Frascati or Oslo Manuals, the survey provides a useful insight into the R&I workforce active within the “Energy and Environment” technology family, which has been selected within the UK government’s innovation strategy as part of a larger group of seven technology classes identified as being of strategic interest. Energy and Environment technologies were the second-most commonly reported technology family in terms of relevance to the role of the surveyed workforce (22%).

The survey also details the proportion of respondents whose work relates to each of the technology families for each sector. Energy and Environment technologies were more prevalent among R&I workers in further education colleges (37%), as well as private sector businesses and public sector research organisations (both 27%) (Figure 2.22, Panel B). With 36%, the Energy and Environment technology domain also ranks among the top three in terms of share of business owners/sole traders (Figure 2.22, Panel A). This suggests a key role for start-ups in R&I in this area.

Figure 2.22. R&I workforce distribution across technology families and sectors

Note: Base is all respondents (7 519). Respondents were able to select multiple options. Percentages in Panel A are weighted.

Source: Government of the United Kingdom (2023[23]), “Insights from the UK-wide survey of the Research and Innovation Workforce 2022”, https://assets.publishing.service.gov.uk/media/641d90305155a2000c6ad5f8/insights-uk-survey-research-innovation-workforce-2022.pdf.

Statlink

Inventors and entrepreneurs

The potential of academic entrepreneurship to transfer valuable scientific knowledge into commercial applications (the application of knowledge is the subject of Chapter 3) is widely recognised by policymakers and economists alike. Indeed, there are various prominent examples among today's most prolific companies that were founded by scientists who came from academia, including Google, DeepMind, BioNTech and Moderna. The unique scientific foundations are frequently assumed to generate innovations with stronger breakthrough potential and high social value (Berger, Dechezleprêtre and Kirpichev Cherezov, forthcoming[24]).

Building on the definition used by Roche, Conti and Rothaermel (2020[25]), an OECD study (Berger, Dechezleprêtre and Kirpichev Cherezov, forthcoming[24]) examines start-ups founded by academic entrepreneurs (AEs), i.e. founders with a doctoral degree. The energy, resources and sustainability sector has the second-highest share of founders with doctoral degrees after the health and biotechnology sector. Start-ups in the media, real estate and travel sectors present the lowest shares of PhD-holding founders (Figure 2.23).

Figure 2.23. Start-ups with PhD-holding founders in OECD countries, by sector, 2000-22

As a percentage of all start-ups in the OECD/STI Start-ups Database

Note: The sample includes 81 318 companies located in 38 OECD countries. The sample size is limited to start-ups with information on the founders’ educational background. Additional restrictions apply to firms that have received any type of funding. This funding must have occurred after 2000; the initial funding must not have occurred before the companies’ founding year and must have happened before a firm’s initial public offering. Sectors are aggregated and harmonised across the two primary databases.

Source: Berger, Dechezleprêtre and Kirpichev Cherezov (forthcoming[24]), Academic Start-ups’ Funding and the Role of Alternative Funding and Support Instruments, based on OECD/STI Start-ups Database.

Statlink

The influence of science on public understanding and policy

Measurement rationale

As noted in Chapter 1, science is indispensable in building broad societal awareness of environmental challenges and their causes. They can also inform an objective and risk-based assessment of options for policy at multiple levels. For example, the accumulation of knowledge about climate change since the 1960s has enabled the scientific community to alert decision makers and the public to the risks of climate change, helping identify avenues for mitigation and their relative costs (Hallegatte et al., 2016[26]). The IPCC has been instrumental and is key in supplying scientific evidence to the United Nations Framework Convention on Climate Change (UNFCCC), an international treaty established in 1992 to co‑ordinate global efforts to combat climate change through mitigation, adaptation and financial support.

Box 2.6. Data sources and methods for the analysis of scientific influence on environmental policy

Science informs and underpins policy at different levels and through many different, hard-to-quantify channels. As a result, policymakers across countries find it difficult to consistently measure the overall uptake of science. However, tracking citation links in systematic scientific compilations, such as in IPCC reports, offers a unique opportunity to gain insight into the nexus between science and policy in the environmental domain. IPCC citations provide an important case to demonstrate the approach by referring to a well-defined, authoritative and transparent scientific consensus mechanism.

While the citations within the latest IPCC report can be comprehensively analysed thanks to structured bibliometric data issued by the IPCC itself, no such data are available for the earlier IPCC reports. To bridge the data gap, this section draws partly on Overton (Szomszor and Adie, 2022[27]), a private data provider that developed a database linking policy documents to the documents they cite and are cited by, by web-crawling publicly accessible documents published by a curated list of over 30 000 organisations. This makes it possible to trace country contributions to Assessment Reports 1‑5 published between 1990 and 2023. Citation parsing is, however, less complete for older documents, and the coverage has been deemed insufficient for Assessment Reports 1 and 2. The database has been used to study also the relevance of climate change research to policy more broadly (Bornmann et al., 2022[28]).

It is still relatively early to make robust comparisons between national policies’ degrees of reliance on science for their environmental and related policies. Explicit recognition of scientific sources is a heterogeneous and evolving practice, dependent on the nature of the policy work, the publication of its outcomes, and transparency on the underlying processes. Scientific influence may not always leave a citation trace, while references may not reflect actual influence.

Indicators of relevance and influence in IPCC reports

The work of the IPCC is organised into three main working groups. Working Group I assesses the physical science basis of climate change, confirming human influence on global warming; Working Group II evaluates climate change impacts, vulnerabilities and adaptation strategies; and Working Group III explores mitigation solutions to reduce GHG emissions. These groups provide a scientific foundation for global climate policies and action under the UNFCCC and Paris Agreement. Findings are published in the form of Assessment Reports, which are released approximately every four years.

In terms of absolute contribution to the most recent 6^th IPCC report, based on fractional counts, the United States leads, followed by the United Kingdom, Australia, Germany and China. Countries have different relative strengths within each working group, however. France, Switzerland and Japan contribute the most, in terms of share of publications, to the physical science basis, ranging from 44% to 38%. South Africa, the Netherlands and Australia contribute the most to Working Group II (72% to 68%), while Denmark, Austria and Sweden contribute the most to Working Group III (37% to 34%) (Figure 2.24).

Figure 2.24. Main contributing countries to scientific publications cited in the 6th IPCC Assessment Report, by working group

Top 20 contributing countries, country of affiliation-based count of scientific publications cited in the IPCC report

Note: Country contributions calculated as fractional counts, with all authors’ affiliations equally weighted, adding up to 1 at the cited article level.

Source: OECD analysis based on electronic bibliographic files provided by the IPCC merged with publication records from OpenAlex.

Statlink

Figure 2.25. Main contributing countries to publications cited in IPCC Assessment Reports 3 to 6, 2001-23

Note: Fractional counts for the ten contributing economies to each Assessment Report, based on publications parsed within the Overton citation database. AR6 results in this figure differ from those displayed in the previous figure. For AR1-AR2, coverage was deemed insufficient. Europe includes France, Germany, the Netherlands, Norway, Sweden, Switzerland and the United Kingdom.

Source: OECD analysis of Overton data merged with OpenAlex scientific publication records.

Statlink

Although the top two contributing countries to the IPCC reports, the United States and the United Kingdom, have remained the same, at least since the third Assessment Report, published in 2001, the relative importance of the United States in the top ten contributing countries to each Assessment Report has declined over time. In the two recent Assessment Reports, China has made an increasing contribution, rising from 4% to 8% (Figure 2.25).

International collaboration plays an important role in the scientific publications underpinning the IPCC reports. In the 6^th Assessment Report, almost one-half of the cited scientific publications are based on single-country authorship, while 34% are based on multilateral collaboration (Figure 2.26). The rate of multilateral collaboration changes across the three working groups, with Working Group III (Mitigation options) relying more extensively on multilateral collaboration compared to Working Group I (Physical science basis) and Working Group II (Vulnerability of socio-economic and natural systems).

Figure 2.26. International collaboration in the 6^th IPCC Assessment Report, by working group

Note: A paper is attributed as relying on a single country if it has multiple authors with affiliations from one single country or a single author and the affiliation of their institution is either known or unknown. It is considered to involve bilateral collaboration if its authors are affiliated with institutions from two different known countries; trilateral or more if its authors are affiliated with institutions from at least three known countries. If there is an additional unknown affiliation in the last case, the article is still considered to belong to the “trilateral or more” category. The international collaboration status of a scientific publication is deemed unknown in all remaining cases.

Source: OECD calculations based on IPCC bib.tex files for AR6 merged with OpenAlex records.

Statlink

A broad range of scientific disciplines contributes to the work of the IPCC. In terms of the thematic distribution of the science cited in the 6^th Assessment Report, “Environmental science” and “Earth and planetary sciences” occupy the top two ranks, while “Agricultural and biological sciences” are also in the top five. “Social sciences” and “Multidisciplinary science”, however, also play an important role in the science base supporting the 6^th IPCC Assessment Report, although the role of the latter appears to be less important in terms of supporting climate change mitigation and adaptation technologies (Figure 2.27).

Figure 2.27. Main scientific fields contributing to the 6^th IPCC Assessment Report

Top 20 ASJC journal fields, fractional counts

Note: ASJC fields are assigned at the journal level, and journals can have one or more ASJC codes assigned to them. In the latter case, each field is assigned an equal fraction, adding to 1 at the article level.

Source: OECD calculation based on information from IPCC bibliographic files matched to Scopus publication records.

Statlink

Inventive activity in energy and environmental technology

Measurement rationale

Given the urgent need to push forward technology boundaries in order to develop environmentally superior goods and services, it is important to understand the dynamics in the technological fields that are aligned with sustainable growth as well as those associated with highly carbon-intensive economic activities. An invention is a unique or novel device, method, composition, idea or process. It can take the form of an improvement upon a machine, product or process to increase efficiency or lower costs. This section focuses on the use of patent statistics as indicators of inventions. Because patents establish claims on what their protected inventions can do and those are “signed off” by experts in patent offices, there is an explicit, objective and direct relevance link to the achievement of energy or environmental goals that is particularly valuable. This explains the significant contribution of these indicators to monitoring STI for sustainable growth (Box 2.7).

Box 2.7. Measuring inventions relevant to the energy and green transition

Patents as indicators of invention

Patents are a means of protecting inventions developed by firms, institutions or individuals, and as such, they may be interpreted as indicators of invention. Before an invention can become an innovation, further entrepreneurial efforts are required to develop it and put it into use or bring it to the market. Patent indicators convey information on the output and processes of inventive activities (OECD, 2009[29]).

Patent related indicators presented in the report are based on patent applications filed under the Patent Co-operation Treaty (PCT) and on patent families:

• A PCT patent application is an international filing process that streamlines entry into multiple jurisdictions.

• Patent families refer to patents filed in different jurisdictions to protect a same invention. Patents that are filed in more than one Industrial Property (IP) office are often considered to be of higher quality because seeking protection in multiple major markets requires significant investment, indicating the applicant's confidence in the patent's value and broad commercial potential. In the case of this publication, IP5 patents are defined as those filed in at least two offices, one of which within the five largest IP offices (Dernis et al., 2015[30]), i.e. the European Patent Office (EPO), the Japan Patent Office (JPO), the Korean Intellectual Property Office (KIPO), the China National Intellectual Property Administration (CNIPA), and the United States Patent and Trademark Office (USPTO).

Measuring the relevance of patents to environmental objectives

Patents provide information on the technological content of the invention (notably its technical field) and the geographical location of the inventive process. The main benefit, in the context of this publication, is that the detailed technological taxonomy in which patent documents are classified, the International Patent Classification (IPC) scheme and the Cooperative Patent Classification (CPC) scheme that extends it, and a number of classification approaches that have been subsequently developed, make it possible to distinguish between innovations based on their environmental attributes.

The two most widely adopted approaches to identifying patented innovations relevant to environmental sustainability are:

• the environment-related technology (ENV-TECH) classification system developed by the OECD (Haščič and Migotto, 2015[31])

• the Y02/Y04S tagging scheme for Climate Change Mitigation and Adaptation Technologies implemented by the European Patent Office (EPO).

The Y02/Y04S patent tagging scheme has several key subclasses that allow for disaggregation by technology families that broadly align with key economic sectors (energy, transport, industry, agriculture, etc.). The related Y04S classification further extends to smart grid technologies and systems that enhance energy distribution and consumption efficiency. The ENV-TECH overlaps with the Y02 definition but is broader.

For monitoring and comparison purposes, the International Energy Agency (IEA) and the EPO have developed an approach to identify “high-carbon” patents (IEA/OECD, 2021[32]), which has been further extended by the OECD within the STI Microdata Lab2.

There is an additional category of “grey” patents, namely those that tend to target efficiency improvements in high-carbon or otherwise polluting technologies, for instance technologies which improve the energy efficiency of the internal combustion engine (Aghion et al., 2016[33]). “Grey” patents definition is adapted from the list of CPC class symbol codes initially published in Dechezleprêtre, Martin and Mohen (2013[34]) and Aghion et al.,(2016[33])3.

Limitations

The limitations of using patents as a proxy for innovation are widely discussed in the literature [e.g. Calel and Dechezleprêtre (2016[35]). Patent filings are not necessarily comparable indicators of inventive activity across different technology domains (a patent in one field may cover a narrower set of claims than in another), and the propensity to patent varies substantially by country and sector, which is why absolute patent counts have limited usefulness as an indicator on its own. Some also argue that patents should not be used as an indicator of R&D success (Reeb and Zhao, 2020[36]); however, due to the features outlined above, patents remain one of the most relevant indicators of R&D output and a key, yet not exclusive, precursor to innovation.

Source: OECD, based on multiple sources, including OECD (2009[29]), Patent Statistics Manual, https://doi.org/10.1787/9789264056442-en.

Patent indicators of inventions with different environmental attributes

Following a period of robust growth between 2000 and 2010, patenting of environment-related and climate-related technologies declined as a share of global patenting (Figure 2.28).

Figure 2.28. Patents in environment-related, low-carbon, grey and high-carbon technologies, 2000-2022

As a percentage of total patent applications filed under the Patent Cooperation Treaty (PCT)

Note: Data refer to families of patent applications filed under the Patent Cooperation Treaty (PCT) by earliest filing date. Data from 2022 onwards are incomplete due to unpublished documents. As described in Box 2.7, the definition of patents on environment technologies largely overlaps with the Y02 definition of climate change mitigation and adaptation technologies. Grey patents tend to target efficiency improvements in high-carbon or otherwise polluting technologies. They have been defined by a list of CPC class symbol codes in Dechezleprêtre, Martin and Mohen (2013[34]) and further elaborated by the OECD.

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed March 2025).

Statlink

This is a signal of relative innovation priorities shifting elsewhere and a potential source of concern, given the urgency of the environmental sustainability transition. The shares of high-carbon and grey patents have remained relatively stable in comparison, though even these have seen a minor decline in share. The decline in the share of sustainability-related patents has been widely commented on in the academic literature (Popp et al., 2020[38]; Martin and Verhoeven, 2022[39]; Aghion et al., 2016[33]), with suggested reasons including energy prices, unfavourable policy changes, the exhaustion of the most promising innovation opportunities or the constrained availability of capital in the aftermath of the 2008 financial crisis. The breakdown of climate change mitigation patents into subcategories reveals that patenting in low-carbon energy technologies has experienced the largest decline in terms of its share of the total, followed by low-carbon technologies in transport and, to a lesser degree, buildings, industry and agriculture (Figure 2.29).

Figure 2.29. Trends in patenting for selected climate change mitigation categories, 2000-2022

As a percentage of total patent applications filed under the Patent Cooperation Treaty (PCT)

Note: Data refer to patent applications filed under the Patent Cooperation Treaty (PCT) by filing date and location of the inventors, using fractional counts. Data from 2022 onwards are incomplete due to unpublished documents.

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed March 2025).

Statlink

The rapid increase of patenting in environment-related technologies in China can be observed both with respect to patents filed under the PCT and with respect to IP5 patent families. With respect to IP5 patent families, China experienced an extraordinary 604.97% increase in patent filings between 2010 and 2020. Korea also showed strong expansion with a 79.06% rise, while the United States (15.84%) and EU27 (12.04%) saw moderate but steady growth. In contrast, Japan (-2.05%) and Germany (-2.19%) experienced slight declines, possibly indicating shifts in their innovation landscapes or a preference for alternative intellectual property strategies (Figure 2.30). The considerable momentum gathered by China in environment-related knowledge creation, demonstrated though patent filings as well as other indicators, deserves further assessment. Recent work by the IFC (2025[15]) shows that while Chinese environmental patents receive high citation counts, their estimated economic value, based on measurement methodology established in Guillard et al. (2021[40]) is somewhat lower compared to those from other middle- or low-income and high-income countries

Figure 2.30. Selected contributing economies to environment-related patenting, 2000-2022 (PCT filings) and 2000-2020 (IP5 patent families)

Note: Data in Panel A refer to patent applications filed under the Patent Cooperation Treaty (PCT) by earliest filing date. Data from 2022 onwards are incomplete due to unpublished documents. Data in Panel B refer to patent families, where patents are filed in at least two offices, one of which is within the group the five largest IP offices, by the earliest filing date and location of the inventors, using fractional counts. Data from 2021 onwards are incomplete due to unpublished documents. In both instances, environmental patents refer to the environment-related technology (ENV-TECH) classification system developed by the OECD (Haščič and Migotto, 2015[31]). The top six countries in terms of absolute counts over the surveyed period are included in both charts.

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed March 2024).

Statlink

The share of environment-related patents ranges between 8.2% in Ireland and almost 25% in Denmark (Figure 2.31). Examining the changes in share over time, the global average remains constant; however, there have been some notable declines in share in Greece, Portugal, Costa Rica and Lithuania. However, it must be noted that given the limited PCT patent portfolios in smaller economies, these changes might be limited only to several patents.

Figure 2.31. Differences and change in environment technology patent intensity across selected countries between 2003-12 and 2013-22

Share of ENV-TECH patents in total, patent applications filed under the Patent Cooperation Treaty (PCT)

Note: Data refer to patent applications filed under the Patent Cooperation Treaty (PCT) by filing date and location of the inventors, using fractional counts. Data from 2022 onwards are incomplete due to unpublished documents.

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed March 2025).

Statlink

Adaptation to climate change is crucial to reduce the risks posed by extreme weather events, rising sea levels, and shifting ecosystems, which threaten lives, infrastructure and economies. Societies can minimise damage and maintain stability by implementing strategies such as resilient infrastructure, sustainable water management, and climate-smart agriculture. The Y02A patent tag is part of the CPC system and covers technologies for climate change adaptation, including water supply, flood protection, resilient infrastructure, and agriculture. It helps track innovation in areas like desalination, drought-resistant crops, and buildings designed for extreme weather. With almost 5% of total patents dedicated to climate change adaptation technology, Norway is in the lead, followed by Chile and Iceland (Figure 2.32). Changes from 2003-12 to 2013-22 remain minor for most countries, although Chile, Mexico and Costa Rica have experienced substantial drops, while Norway, Colombia and Latvia have seen an increase.

The content of climate change relevant patenting varies substantially among technology domains. Electrical machinery, engines, pumps and turbines, and environmental technology comprise over 30% of climate change relevant patents, followed by materials, thermal devices, and transport. The share of patents in the engines, pumps and turbines, environmental technology, and thermal devices dropped by several percentage points between 2008-12 and 2018-22 (Figure 2.33).

Figure 2.32. Differences and change in patenting intensity for climate change adaptation technologies across selected countries between 2003-12 and 2013-22

Share of Y02A patents in total, patent applications filed under the Patent Cooperation Treaty (PCT)

Note: Y02A patents are classified under the Y02 section of the Cooperative Patent Classification (CPC) system, which focuses on climate change mitigation and adaptation technologies. Specifically, the Y02A category covers technologies for adaptation to climate change; these are inventions designed to help societies and ecosystems cope with the effects of climate change. Data refer to patent applications filed under the Patent Cooperation Treaty (PCT) by filing date and location of the inventors, using fractional counts.

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed March 2025).

Statlink

Figure 2.33. Share of climate change mitigation and adaptation technologies across WIPO domains, 2008-12 and 2018-22

Note: Data refer to patent applications filed under the Patent Cooperation Treaty (PCT) by priority date and technology field, using fractional counts. Data from 2022 onwards are incomplete due to unpublished documents. Patents are allocated into 35 technology domains, as identified by the World Intellectual Property Organisation’s (WIPO) concordance between International Patent Classification (IPC) codes and technology areas (Schmoch, 2008[41]). Climate change mitigation and adaptation patents are identified using the Y02 tag of the Cooperative Patent Classification (CPC).

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed March 2025).

Statlink

Science for green technology

Measurement rationale

There is broad agreement on the importance of scientific knowledge as a foundation for major technological advances, and on the need for new insights into fundamental principles of matter and energy. Yet, evidence remains limited on the specific contribution of science and scientific institutions to breakthrough innovations. Given the constraints of current technologies, there is significant scope (Box 2.8) to examine how scientific research underpins inventions relevant to environmental sustainability. A clearer understanding of how energy and environmental technologies build on basic and applied science would provide valuable guidance for policymakers.

Box 2.8. Measuring the relevance of science for climate change related technology

Patents citing non-patent literature as traces of science's impact on technology

Prior knowledge on which patented inventions rely may encompass patents, scientific work and other sources. Information contained in references to scientific literature, conference proceedings, databases, and other relevant literature, known as non-patent literature (NPL), can shed light on the knowledge flows between science and innovation, as well as on the scientific pillars of technology. NPL referenced by inventors and patent examiners includes scientific literature as well as other literature that reflects the state of the art in an area, such as trade journals, standards, etc. Tracing which of those references link to scientific publications makes it possible to characterise the science considered relevant for patented inventions.

With some limitations, such as the fact that there may be reasons for citing science beyond the recognition of relevant prior knowledge, citations in patent documents to NPL offer an opportunity to trace the impact of scientific research (van Raan, 2017[42]) and gather valuable insights into the differences in reliance on science between patents with different environmental attributes (Perrons, Jaffe and Le, 2020[3]). Research suggests that there is a correlation between the quality of science cited in a patent and its value, with patents that ranked high on the quality of linked science being twice as valuable as patents linked to low-quality science, which, in turn, are about as valuable as patents without a direct science link (Poege et al., 2019[43]).

Data sources and methods

The indicators in this section rely on the “Reliance on science” dataset, which offers an open-access compilation of NPLs from the front pages or the text of patents granted globally to scientific papers (Marx and Fuegi, 2020[44]; 2021[45]). This dataset offers a comprehensive curated linkage between NPLs and scientific literature, encompassing a collection of approximately 16.8 million citations extracted from the full text of patents granted by the USPTO and the EPO. These 16.8 million citations initially represent a broad pool identified through automated matching algorithms, extracted from the patent’s front page, text or both, and each is assigned a confidence score that reflects the accuracy of the patent-to-article linkage. Citations with the highest confidence scores indicate very high reliability of the match. To ensure methodological robustness, recall performance has been assessed using a carefully designed validation procedure involving 5 939 randomly selected, cross-verified “known good” citations. These citations represent a high-confidence benchmark that the dataset creators have never encountered, ensuring an unbiased recall assessment. At a precision threshold of 99.4%, the recall rate is 78% for the complete evaluation set. Moreover, recall performance increases to 88% when focusing specifically on patent citations originally specified without errors or inconsistencies. The sample was constrained to citations listed explicitly on the front pages of patents, and selecting only those matches with a confidence score higher than five resulted in a refined subset. This filtering reduced the dataset from the original 16.8 million citations to 5.7 million records, of which 98.2% successfully matched with extracted complete records from the OpenAlex database. With that connection, further information can be extracted about the scientific publications cited and their authors’ affiliations, either from within OpenAlex or other bibliographic databases linked through the publication’s digital object identifiers (DOIs).

The process of linking patents to scientific literature is fraught with challenges, however, primarily due to the unstructured nature of patent citations, which often contain inconsistencies such as misspellings or incomplete information. While not all NPL is peer-reviewed science, a significant share is, and therefore, this component can be considered suggestive of potential spillovers from science to private and public R&D efforts. For all technologies, the trend is towards greater reliance on NPL in patented innovations over time, although this can also reflect a change in examination practices. Given its high potential, work on this newly matched database is ongoing as a joint endeavour of the OECD Committees for Scientific and Technological Policy (CSTP) and Industry Innovation and Entrepreneurship (CIIE).

Green inventions’ reliance on knowledge sources other than patents

Climate change mitigation and adaptation patents rely on non-patent literature (NPL) significantly more than high-carbon patents and “grey” patents (i.e. those that improve the efficiency of high-carbon technologies) (Figure 2.34). This result is also echoed in the literature, albeit for a smaller sample of patents constrained to energy (Persoon, Bekkers and Alkemade, 2020[46]).

Figure 2.34. Patents citing non-patent literature, by climate change related attributes, 2007-22

As a percentage of all patents in each patent technology category

Note: Data refer to patent applications filed under the Patent Cooperation Treaty (PCT) by filing date. Patents in climate change mitigation and adaptation technologies, grey technologies, and high-carbon technologies are identified and classified based on Cooperative Patent Classification (CPC) codes and keyword searches in the patent document. See previous section on inventive activity (Box 2.7) for details. Patents citing scientific publications are those that contain a non-patent reference to scientific articles, chemical abstracts, or biological abstracts. Patents relating to climate change (positively or negatively) are overall less likely to cite the NPL than the average patented invention, a body that is dominated in most countries by ICT and pharmaceutical patents.

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed June 2025).

Statlink

Climate change mitigation and adaptation patents tend to have a higher intensity of citation of NPL compared to high-carbon and grey patents across most countries, except Denmark (Figure 2.35). It is important to note, however, that patents relating to climate change (positively or negatively) are overall less likely to cite the NPL than the average patented invention, a body that is dominated in most countries by ICT and pharmaceutical patents. It is not particularly meaningful to compare such very different technology domains, but it is more informative to focus on comparing within domains with an energy technology orientation but with different environmental implications, as discussed in the definitions provided in Box 2.7.

Figure 2.35. Patents citing non-patent literature, by climate change related attributes, in selected economies, 2017-21

As a percentage of all domestic patents, by technology and inventor’s location, selected economies

Note: Data refer to patent applications filed under the Patent Cooperation Treaty (PCT) by filing date and inventor location, using fractional counts. Patents in climate change mitigation and adaptation, “grey”, and high-carbon technologies are identified based on the Cooperative Patent Classification (CPC) codes and keyword searches in the patent document. A minimum reporting threshold of 100 patents per technology area applies. Patents citing scientific publications are those that contain a non-patent reference to scientific articles, chemical abstracts, or biological abstracts. Patents relating to climate change (positively or negatively) are overall less likely to cite the NPL than the average patented invention, a body that is dominated in most countries by ICT and pharmaceutical patents.

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed July 2024).

Statlink

Reliance on NPL in climate change mitigation and adaptation patents increased across most countries between 2007-11 and 2017-21 (Figure 2.36). This might suggest an increasing shift towards “deep tech” technologies within this subset, which could be expected to be more dependent on science. The possibility of the “low-hanging fruit” already being picked and marginal low-carbon innovation requiring more effort (and possibly deeper scientific foundations) has been raised as a possible reason for the decline in low-carbon patent share observed in Figure 2.29 (Popp et al., 2020[38]).

Figure 2.36. Climate change mitigation and adaptation patents citing non-patent literature, in selected economies, 2007-11 and 2017-21

As a percentage of each economy’s climate change mitigation and adaptation patents, by inventor’s location

Note: Data refer to patent applications filed under the Patent Cooperation Treaty (PCT) by filing date and location of the inventors, using fractional counts. Patents in low-carbon technologies are based on the Cooperative Patent Classification (CPC). Only economies with more than 100 patents in low-carbon technologies are included. Patents citing scientific publications are those that contain a non-patent reference to scientific articles, chemical abstracts, or biological abstracts.

Source: OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed July 2024).

Statlink

Scientific foundations of climate change relevant inventions

Climate change mitigation and adaptation technological innovation is driven by diverse scientific disciplines, often emerging at the intersection of multiple fields. New OECD analysis shows that while there are similarities between low-carbon, high-carbon and grey (i.e. ambiguous) technology patents (see Box 2.7 for definitions) with respect to the scientific domains they tend to cite, there are some important differences.

Inventions on technologies that help address environmental challenges build on science to a greater extent than inventions enhancing existing polluting technologies. Although new low-carbon technology patents rely only marginally more on NPL, as shown in the previous subsection, they account for nearly six times as many publications in their citations as high-carbon patents and those with ambiguous carbon-emission impacts. The implication is that the latter two rely relatively more on trade literature, a possible indication of differences in technology maturity.

In addition to engineering (17%), the key fields most cited by low-carbon patents are chemistry (15%) and materials science (12%). As a share of all citations, computer science (7%) is 5 percentage points more important for low-carbon than for high-carbon patents (2%) (Figure 2.37). The wide-ranging nature of these scientific influences underscores the challenge of pinpointing a single dominant field driving the development of new low-carbon technologies and associated innovations.

Examining the contribution of individual countries to the scientific basis that underpins climate change-related patents reveals that nearly 40% of scientific publications cited in low-carbon patents are by US-based authors, followed by China with 13% and Germany and Japan with 8% each. This distribution is expected to change as cited publications by authors based in China are four years more recent than the average (Figure 2.38).

Figure 2.37. Scientific papers cited in climate change mitigation and adaptation, high-carbon and ‘grey’ technology patents, by journal fields, 2015-20

Percentage of articles cited by patents by ASJC journal fields

Note: Data refer to scientific articles cited in green patents filed at the European Patent Office (EPO) or at the United States Patent and Trademark Office (USPTO). Patents in climate change related technologies are identified using the Y02 tag of the Cooperative Patent Classification (CPC). Fractional counts by ASJC are applied. Linkages were established using the patent-to-paper citations dataset produced by Marx and Fuegi (2020[44]).

Source: OECD, calculations based on Scopus Custom Data, Elsevier, Version 1.2024 and OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed July 2024).

Statlink

Figure 2.38. Share of scientific publications cited in climate change mitigation and adaptation patents by author country, 2006-13

As a percentage of the total publications, based on whole counts

Note: Data refer to scientific articles cited in climate change mitigation and adaptation patents filed at the European Patent Office (EPO) or at the United States Patent and Trademark Office (USPTO). Patents in climate change related technologies are identified using the Y02 tag of the Cooperative Patent Classification (CPC). Fractional counts by ASJC are applied. Linkages were established using the patent-to-paper citations dataset produced by Marx and Fuegi (2020[44]).

Source: OECD, calculations based on Scopus Custom Data, Elsevier, Version 1.2024 and OECD (n.d.[37]), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats (accessed July 2024).

Statlink

Business R&D investment related to energy and the environment

Measurement rationale

The previous sections, through their analysis of inventive activity and the R&I workforce, have touched on the contribution of businesses to generating knowledge related to achieving the energy and environmental SDGs. In market and mixed economies, businesses are key actors in R&I systems, developing and implementing new ideas to pursue economic opportunities. The business sector's role as a driver of total R&D expenditure is becoming increasingly important qualitatively and quantitatively (OECD, 2023[47]). However, there is considerable uncertainty as to how much of this growth has the potential to contribute to energy and environmental goals.

Policymakers are equally interested in ascertaining whether business R&D efforts result in incremental improvements in existing technologies or drive breakthroughs that potentially allow for new industries to develop. It is often presumed that, compared to R&D performed by government or non-profit institutions, business R&D is more results-driven, with a strong focus on immediate application, profitability and shareholder value. Such a statement may not apply equally to all types of business. To assess these issues, it is important to look not only at inventive outcomes but also to assess the resources businesses dedicate to such goals and how different types of businesses differ in their approaches.

Box 2.9. Measuring R&D relevant to energy and environmental goals in business surveys

The main aggregate statistic used to describe R&D performance within the business enterprise sector is business enterprise R&D (BERD), the component of gross domestic expenditure on R&D (GERD) incurred by units in the business enterprise sector. Several countries measure R&D as relevant to energy and environmental goals but do so differently, preventing consolidated international comparisons. In their official business R&D surveys, few countries (e.g. Portugal) currently ask respondents to distribute their R&D expenditure across socio-economic objective (SEO) categories (see Chapter 4), such as energy and the environment. In addition to the need to keep surveys short, one key reason is that businesses do not necessarily view their R&D activities in terms of a classification system conceived for allocating government expenditures and funding.

OECD guidance in the Frascati Manual recognises that countries can find it useful to obtain information on the relevance of business sector R&D towards specific societal goals, but given thematic overlaps and diversity of priorities, suggest that this is not done on a mutually exclusive basis but through ad hoc questions (OECD, 2015[48]; Galindo-Rueda and López-Bassols, 2022[49]). Although there is no international reporting consensus at present, Box 2.10 presents examples of such practices for Canada and Norway. France’s business R&D survey collects information on R&D for three main resource and environmental goals but, similarly to SEO classifications, requests that the reported amounts be mutually exclusive.

Distributing R&D to economic activities (i.e. industries) is a possible alternative. However, it presents several challenges, starting from the point that economic activity is a rather ambiguous proxy for energy and environmental performance relevance. For example, a company in the R&D professional services sector may conduct R&D for an energy supplier company, or an equipment manufacturer may conduct R&D to improve the energy efficiency of its products or internal production processes. Industries are heterogeneous in their energy and environmental orientation for R&D and turnover. In none of these cases would the R&D expenditure be, based on the firm’s principal industrial activity, on the energy sector (“main activity” approach). Only the former case would be allocated to energy if a user or “industry-orientation” approach were followed. The Frascati Manual 2015 recommends classifying statistical units by their main activity and suggests that their R&D activities be classified both on that basis and by industry orientation.

Business R&D on energy and the environment

R&D in the energy, water and waste management utility industries

The proportion of business R&D companies conduct in the “electricity, gas and water supply; sewerage, waste management and remediation activities” is the highest in Latvia, Estonia and Portugal (Figure 2.39).

Figure 2.39. Business expenditure on R&D in the utilities and waste management industries, selected countries, 2012 and 2022

Note: Figures are based on estimates of business R&D by industry reported on a main activity basis, in ISIC Rev.4.The utilities and waste management industry correspond to the ISIC Rev.4 Sections D (Electricity, gas, steam and air conditioning supply) and E (Water supply; sewerage, waste management and remediation activities). These statistics are based on OECD R&D Statistics (http://oe.cd/rds) and ANBERD (http://oe.cd/anberd) Databases. Data refer to 2012 and 2022 except for Australia (2012, 2021), Austria (2012, 2021), Belgium (2012, 2021), Canada (2012, 2021), Chile (2012, 2021), Croatia (2012, 2021), Denmark (2012, 2020), France (2012, 2021), Germany (2012, 2021), Ireland (2012, 2021), Israel (2013, 2021), Latvia (2013, 2020), the Netherlands (2013, 2021), New Zealand (2021), Poland (2010, 2022), Sweden (2012, 2021) and the United States (2012, 2020 – for which only range estimates are available to prevent confidential data disclosure).

Source: OECD (n.d.[50]), ANBERD Database, http://oe.cd/anberd (accessed November 2024).

Statlink

Estonia and Lithuania experienced a substantial decline in the share of business R&D from 2012 to 2022. The United States, Romania and Israel dedicate the smallest share of business R&D to these industries. US (data expressed as interval) and French companies were the largest R&D performers in 2021 among countries for which data are available, including all major OECD countries but not China.

Box 2.10. Measurement case studies: Canada and Norway’s business R&D surveys

Statistics Canada’s estimates of business R&D on clean technologies

Statistics Canada uniquely collects data on energy R&D expenditures disaggregated by area of technology as part of its Annual Survey of Research and Development in Canadian Industry. This approach follows IEA guidance, also reflected in OECD R&D measurement proposals, and allows for a detailed overview of the investments businesses make into relevant R&D. In 2022, businesses spent CAD1.9 billion (Canadian dollars) on in-house R&D to further develop “clean” (i.e. low-carbon) technologies. Energy efficiency saw the strongest growth in in-house R&D spending. This was followed by hydrogen and fuel cells, and electric power.

The percentage share of total in-house energy-related R&D for each of these three technology areas has also grown considerably since 2014. For instance, from 2014 to 2022, energy efficiency increased from 5.8% to 21.5%; hydrogen and fuel cells, from 3.6% to 11.1%; and electric power, from 3.7% to 13.5%. In-house R&D expenditures on fossil fuels also increased for the second consecutive year, up by CAD 171 million from 2021 to CAD 869 million (+24.5%) in 2022 (Figure 2.40).

Figure 2.40. In-house energy-related R&D expenditures on clean technologies, 2010-22

In CAD millions

Source: OECD, based on Statistics Canada (2024[51]), Energy-related research and development expenditures, 2022, https://www150.statcan.gc.ca/n1/daily-quotidien/240919/dq240919b-eng.htm.

Statlink

Statistics Norway’s R&D in the business enterprise sector by thematic area

Statistics Norway similarly collects detailed data on R&D in the business sector, including by detailed technology thematic areas. The thematic areas are defined in sufficiently detailed terms to allow for the evaluation of trends with respect to business R&D in sustainable vs high-carbon energy, as well as environmental themes beyond energy. The petroleum thematic area saw a decline in business R&D in absolute terms between 2015 and 2022 by NOK 685 million (Norwegian kroner), while the sustainable energy and environmental sustainability-related thematic areas saw substantial increases. Businesses spent almost NOK 1.4 billion more on R&D in the energy efficiency and change theme (formerly energy efficiency), for instance, and almost NOK 1.5 billion more on environmental technology. The renewable energy and climate technology and other emission reductions thematic areas have also registered significant gains in business R&D spend.

Figure 2.41. Norway’s Business Expenditure on R&D by selected thematic area, 2015 and 2022

Current BERD cost (NOK millions)

Source: OECD, based on Statistics Norway (n.d.[52]), 11483: Thematic area of R&D in the business enterprise sector. Current cost (NOK million), by contents, industry (SIC2007) and year, https://www.ssb.no/en/statbank/table/11483/.

Statlink

R&D on energy

R&D in the business sector contributing to energy applications is significantly higher than R&D conducted by energy and other utilities, which account, on average, for less than 2% of the total. In the United States, where energy and environmental R&D questions have been used systematically for one decade, nearly 6% of business R&D is oriented to energy applications and 2% to environmental protection.

As shown in Figure 2.42, the energy utilities, extraction and petroleum and coal-product-transforming industries show the highest energy R&D intensity, i.e. percentage of energy-oriented R&D as a percentage of total industry’s R&D. However, these are not the largest contributors in absolute levels to energy R&D. Because of their much larger contribution to total R&D, the semiconductor and electronic component industry and the motor vehicle are the main investors in energy R&D, with shares close to 20% of total R&D.

New indicators of R&D investment by firms in selected markets and technology areas

As shown throughout this report, using existing data sources and classifications results in a reasonable but incomplete characterisation of the relevance of STI activities to energy and environmental goals, driving the need for new complementary data sources and indicators. This is also the case for business R&D, where the challenge stems from the heterogeneity of business activities and R&D portfolios and disclosure limits posed by the need to preserve assurances of confidentiality when collecting statistical data from firms.

The OECD Short-term Financial Tracker of Business R&D (SwiFTBeRD) dashboard helps monitor the latest trends in R&D investments by the world's top R&D business investors based on public annual and quarterly company reports (Box 2.11). For regulatory purposes, R&D disclosures are mandatory for large companies active in specific financial markets, and investors have come to expect regular reporting on those. While the motivation for exploring company reports stemmed originally from the limited timeliness of official statistics that results from annual or even biennial collection and reporting cycles, another set of possible insights from such data sources results from the qualitative information that is contained in company reports about the main lines of business of a company.

Figure 2.42. Business R&D expenditure in the energy application area across US industries, 2019

Total expenditure and as a percentage of each industry’s R&D expenditure

Note: The number of companies that remained within the scope of the survey between sample selection and tabulation was 42,500.

Source: OECD based on National Center for Science and Engineering Statistics (NCSES). 2020. Business Enterprise Research and Development: 2019. NSF 22-329. Alexandria, VA: National Science Foundation. Available at https://ncses.nsf.gov/pubs/nsf22329/.

Statlink

Box 2.11. Measuring the latest business R&D trends using corporate reports

The OECD Short-term Financial Tracker of Business R&D (SwiFTBeRD) dashboard allows users to visualise quarterly, semi-annually and annually reported R&D data for the world's top R&D investors, providing company-specific and sectoral insights. It aims to deliver the timeliest possible view of R&D data reported by companies, with updates published continuously, shortly after they have been released in their quarterly financial reports. The initiative seeks to address the growing demand for real-time, high-quality R&D data to support evidence-based policy making, particularly in areas like innovation, economic growth and industrial strategy.

Company reports of R&D expenses need not coincide with R&D expenditures as covered in official R&D statistics compiled according to the Frascati Manual (OECD, 2015[53]). In order to compile the data presented in the SwiFTBeRD Dashboard, the OECD implements a series of adjustments to enhance comparability whenever the necessary information is available. Companies presenting their financial results in compliance with the International Financial Reporting Standards capitalise part of their development costs (under some criteria). In the data presented in SwiFTBeRD, capitalised development costs are added to reported R&D expenses, while amortisation of capitalised development expenditures is conversely excluded, provided that the information is available both in the annual and interim reports. In addition, when possible, expenses and impairment of purchased in-process R&D (as well as restructuring R&D costs) are excluded in the SwiFTBeRD figures so as to align as much as possible with R&D conducted in the reference period and deliver more meaningful indicators.

For the purposes of this publication, the OECD SwiFTBERD panel has been extended to include leading companies from the established alternative energy sector, which includes companies active in low-carbon energy technologies, such as Siemens Energy, Vestas Wind Systems, SolarEdge Technologies, Sungrow Power Supply, Orano, First Solar, SMA Solar Technology and Nordex. The electricity sector data include Electricité de France, Korea Electric Power, CGN Power, Iberdrola, GCL Technology, Energias de Portugal, ChargePoint, Landis+Gyr and Tokyo Electric Power.

Source: OECD (2025[54]), OECD SwiFTBeRD Dashboard, https://oecd-main.shinyapps.io/swiftberd/.

For the purposes of this publication and future monitoring, the OECD SwiFTBERD panel has been recently extended to include leading companies from the established alternative energy and electricity sectors. Analysis of trends for those companies shows robust growth in the R&D expenditure in this sector, with the growth since 2017 being second only to the software, computer and electronics sector (Figure 2.43). This performance, which is also matched by growth in revenue, needs to be considered in the context of ongoing supply chain difficulties faced by some of the companies in the group, as well as intensive competition from China, for which it is more difficult to obtain comparable data. The gap in R&D trends in real terms between alternative energy and electricity is increasing over time.

Figure 2.43. Trends in R&D and revenue for the SwiFTBERD panel of R&D investors, by industry groups, 2017-24

Note: The alternative energy sector data include Siemens Energy, Vestas Wind Systems, SolarEdge Technologies, Sungrow Power Supply, Orano, First Solar, SMA Solar Technology and Nordex. The electricity sector data include Electricité de France, Korea Electric Power, CGN Power, Iberdrola, GCL Technology, Energias de Portugal, ChargePoint, Landis+Gyr and Tokyo Electric Power. For the other sectors, the company coverage is detailed in the OECD SwiFTBeRD Dashboard (https://oecd-main.shinyapps.io/swiftberd/). Annual data correspond to calendar year data. For companies whose fiscal year-end is November, December or January, figures are based on R&D reported in annual accounts. For Tokyo Electric Power and Landis+Gyr, annual data cover the period from March to February of the following year. For the other companies, figures are obtained by recombining successive quarterly reports' data. Reported values are deflated using the gross domestic product (GDP) price index of OECD countries.

Source: OECD calculations, based on OECD (2025[54]), SwiFTBeRD Dashboard, https://oecd-main.shinyapps.io/swiftberd/ (accessed April 2025), and data extracted from annual reports for alternative energy and electricity companies.

Statlink

References

[33] Aghion, P. et al. (2016), “Carbon Taxes, Path Dependency, and Directed Technical Change: Evidence from the Auto Industry”, Journal of Political Economy, Vol. 124/1, pp. 1-51, https://doi.org/10.1086/684581.

[14] Aristodemou, L. et al. (forthcoming), Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods.

[24] Berger, M., A. Dechezleprêtre and D. Kirpichev Cherezov (forthcoming), Academic Start-ups’ Funding and the Role of Alternative Funding and Support Instruments, OECD Publishing, Paris.

[28] Bornmann, L. et al. (2022), “How relevant is climate change research for climate change policy? An empirical analysis based on Overton data”, PLOS ONE, Vol. 17/9, p. e0274693, https://doi.org/10.1371/journal.pone.0274693.

[8] Bush, V. (1945), Science: The Endless Frontier.

[35] Calel, R. and A. Dechezleprêtre (2016), “Environmental policy and directed technological change: Evidence from the European carbon market”, Review of Economics and Statistics, Vol. 98/1, pp. 173-191, https://doi.org/10.1162/rest_a_00470.

[10] Chapin, D., C. Fuller and G. Pearson (1954), “A New Silicon <i>p-n</i> Junction Photocell for Converting Solar Radiation into Electrical Power”, Journal of Applied Physics, Vol. 25/5, pp. 676-677, https://doi.org/10.1063/1.1721711.

[9] Dalla Fontana, S. and R. Nanda (2023), “Innovating to net zero: Can venture capital and start-ups play a meaningful role?”, Entrepreneurship and Innovation Policy and the Economy, Vol. 2, pp. 79-105, https://doi.org/10.1086/723236.

[34] Dechezleprêtre, A., R. Martin and M. Mohen (2013), Knowledge Spillovers from Clean and Dirty Technologies: A Patent Citation Analysis.

[30] Dernis, H. et al. (2015), World Corporate Top R&D Investors: Innovation and IP bundles, https://doi.org/10.2791/741349.

[6] European Commission (2024), Science, research and innovation performance of the EU, 2024 – A competitive Europe for a sustainable future, https://data.europa.eu/doi/10.2777/965670.

[21] Eurostat (n.d.), Students enrolled in tertiary education by education level, programme orientation, sex and field of education, https://ec.europa.eu/eurostat/databrowser/view/educ_uoe_enrt03__custom_15308082/default/table?lang=en&page=time:2021 (accessed on 11 March 2025).

[49] Galindo-Rueda, F. and V. López-Bassols (2022), “Implementing the OECD Frascati Manual: Proposed reference items for business R&D surveys”, OECD Science, Technology and Industry Working Papers, No. 2022/03, OECD Publishing, Paris, https://doi.org/10.1787/d686818d-en.

[22] Government of France (2024), Thèses soutenues en France depuis 1985, https://www.data.gouv.fr/fr/datasets/theses-soutenues-en-france-depuis-1985/ (accessed on 11 March 2025).

[23] Government of the United Kingdom (2023), “Insights from the UK-wide survey of the Research and Innovation Workforce 2022”, BEIS/DSIT Research Paper, No. 2023/004, https://assets.publishing.service.gov.uk/media/641d90305155a2000c6ad5f8/insights-uk-survey-research-innovation-workforce-2022.pdf.

[11] Green, M. (2000), “Photovoltaics: technology overview”, Energy Policy, Vol. 28/14, pp. 989-998, https://doi.org/10.1016/s0301-4215(00)00086-0.

[40] Guillard, C. et al. (2021), Efficient industrial policy for innovation: standing on the shoulders of hidden giants, CEP Discussion Papers, https://eprints.lse.ac.uk/113832/.

[26] Hallegatte, S. et al. (2016), “Mapping the climate change challenge”, Nature Climate Change 2016 6:7, Vol. 6/7, pp. 663-668, https://doi.org/10.1038/nclimate3057.

[31] Haščič, I. and M. Migotto (2015), “Measuring environmental innovation using patent data”, OECD Environment Working Papers, No. 89, OECD Publishing, Paris, https://doi.org/10.1787/5js009kf48xw-en.

[32] IEA/OECD (2021), Methodology for identifying fossil fuel supply related technologies in patent data, https://iea.blob.core.windows.net/assets/d14427c6-2aa2-4422-9074-5a68940a5a96/Methodology_for_identifying_fossil_fuel_supply_related_technologies_in_patent_data.pdf.

[15] IFC (2025), Innovation in Green Technologies: Opprotunity for Businesses in Low and Middle-income Countries?, https://www.ifc.org/content/dam/ifc/doc/2025/innovation-in-green-tech.pdf.

[12] IPCC (2023), Climate Change 2021 – The Physical Science Basis, Cambridge University Press, https://doi.org/10.1017/9781009157896.

[39] Martin, R. and D. Verhoeven (2022), The Private Value of Clean Energy Innovation.

[45] Marx, M. and A. Fuegi (2021), “Reliance on science by inventors: Hybrid extraction of in‐text patent‐to‐article citations”, Journal of Economics and Management Strategy, Vol. 31/2, pp. 369-392, https://doi.org/10.1111/jems.12455.

[44] Marx, M. and A. Fuegi (2020), “Reliance on science: Worldwide front‐page patent citations to scientific articles”, Strategic Management Journal, Vol. 41/9, pp. 1572-1594, https://doi.org/10.1002/smj.3145.

[13] Ministry of Higher Education and Science (2022), Rumbaseret grøn forskning – Bibliometrisk analyse af Danmarks rumbaserede grønne forskning September 2022, http://www.ufm.dk.

[54] OECD (2025), OECD SwiFTBeRD Dashboard, https://oecd-main.shinyapps.io/swiftberd/.

[17] OECD (2024), OECD Employment Outlook 2024: The Net-Zero Transition and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/ac8b3538-en.

[18] OECD (2024), OECD Employment Outlook 2024: The Net-Zero Transition and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/ac8b3538-en.

[47] OECD (2023), “The Impact of R&D tax incentives: Results from the OECD microBeRD+ project”, OECD Science, Technology and Industry Policy Papers, No. 159, OECD Publishing, Paris, https://doi.org/10.1787/1937ac6b-en.

[7] OECD (2021), OECD International Survey of Science, http://oe.cd/issa.

[16] OECD (2017), OECD Science, Technology and Industry Scoreboard 2017: The Digital Transformation, OECD Publishing, Paris, https://doi.org/10.1787/9789264268821-en.

[53] OECD (2015), Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental Development, The Measurement of Scientific, Technological and Innovation Activities, OECD Publishing, Paris, https://doi.org/10.1787/978926.

[48] OECD (2015), Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental Development, The Measurement of Scientific, Technological and Innovation Activities, OECD Publishing, Paris, https://doi.org/10.1787/9789264239012-en.

[29] OECD (2009), Patent Statistics Manual, OECD Publishing, Paris, https://doi.org/10.1787/9789264056442-en.

[50] OECD (n.d.), ANBERD Database, http://oe.cd/anberd.

[20] OECD (n.d.), Research and Innovation Careers Observatory, https://www.oecd.org/en/networks/research-and-innovation-careers-observatory.html.

[5] OECD (n.d.), Science and bibliometric indicators, https://www.oecd.org/en/data/datasets/science-and-bibliometric-indicators.html.

[37] OECD (n.d.), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats.

[4] OECD and SCImago Research Group (CSIC) (2016), Compendium of Bibliometric Science Indicators, OECD, Paris, https://web-archive.oecd.org/2016-09-29/415063-Bibliometrics-Compendium.pdf.

[3] Perrons, R., A. Jaffe and T. Le (2020), “Tracing the linkages between scientific research and energy innovations: A comparison of clean and dirty technologies”, NBER Working Paper Series, http://www.nber.org/papers/w27777.

[46] Persoon, P., R. Bekkers and F. Alkemade (2020), “The science base of renewables”, Technological Forecasting and Social Change, Vol. 158, p. 120121, https://doi.org/10.1016/j.techfore.2020.120121.

[43] Poege, F. et al. (2019), “Science quality and the value of inventions”, Science Advances, Vol. 5/12, https://doi.org/10.1126/sciadv.aay7323.

[38] Popp, D. et al. (2020), “Innovation and entrepreneurship in the energy sector”, NBER Working Paper Series, No. 27145, https://doi.org/10.3386/W27145.

[36] Reeb, D. and W. Zhao (2020), “Patents do not measure innovation success”, Critical Finance Review, Vol. 9/1-2, pp. 157-199, https://doi.org/10.1561/104.00000087.

[25] Roche, M., A. Conti and F. Rothaermel (2020), “Different founders, different venture outcomes: A comparative analysis of academic and non-academic startups”, Research Policy, Vol. 49/10, p. 104062, https://doi.org/10.1016/j.respol.2020.104062.

[2] Rogelj, J. (2023), “Net zero targets in science and policy”, Environmental Research Letters, Vol. 18/2, https://doi.org/10.1088/1748-9326/acb4ae.

[41] Schmoch, U. (2008), Concept of a Technology Classification for Country Comparisons, http://193.5.93.80/edocs/mdocs/classifications/en/ipc_ce_41/ipc_ce_41_5-annex1.pdf.

[51] Statistics Canada (2024), Energy-related research and development expenditures, 2022, https://www150.statcan.gc.ca/n1/daily-quotidien/240919/dq240919b-eng.htm.

[52] Statistics Norway (n.d.), “11483: Thematic area of R&D in the business enterprise sector. Current cost, by detailed industry (SIC2007) (NOK million) 2015-2023”, Research and development in the business enterprise sector, https://www.ssb.no/en/statbank/table/11483/.

[27] Szomszor, M. and E. Adie (2022), “Overton: A bibliometric database of policy document citations”, Quantitative Science Studies, Vol. 3/3, pp. 624-650, https://doi.org/10.1162/qss_a_00204.

[19] UNESCO (2015), International Standard Classification of Education.

[42] van Raan, A. (2017), “Patent citations analysis and its value in research evaluation: A review and a new approach to map technology-relevant research”, Journal of Data and Information Science, Vol. 2/1, pp. 13-50, https://doi.org/10.1515/jdis-2017-0002.

[1] World Meteorological Organisation (2023), United in Science 2023, https://library.wmo.int/idurl/4/68235.

Notes

← 1.  The “planet” umbrella comprises: SDG 6: Clean water and sanitation; SDG 12: Responsible consumption and production; SDG 13: Climate action; SDG 14: Life below water; and SDG 15: Life on land.

← 2.  High carbon patents include fossil fuels patents as identified in (IEA/OECD, 2021[32]) + patents in the following CPC - F02, C21B5, C21B7, C21B9, C10G1, C10L1, C10J, E02B, F01K, F02C, F22, F23, F27, F28, F24J, C01F7, C03B5, C04B7, C04B9, C04B11, C05C, C05D, C01C1, C10, C21B5, C21B7, C21B9, D21C3 ) but not among CPC Y02 patents

← 3.  Grey technologies refer to fossil fuels patents as identified in (IEA/OECD, 2021[32]) + patents in the following CPC - F02, C21B5, C21B7, C21B9, C10G1, C10L1, C10J, E02B, F01K, F02C, F22, F23, F27, F28, F24J, C01F7, C03B5, C04B7, C04B9, C04B11, C05C, C05D, C01C1, C10, C21B5, C21B7, C21B9, D21C3 ) AND patents among CPC Y02 patents, plus patents in Y02E20, Y02T10/12, Y02T10/40, Y02P10/25, Y02P20/129, Y02P40/121, Y02P40/50, Y02P80/1

Click to read more