OECD Science, Technology and Innovation Outlook 2018

Chapter 1. An introduction to the STI Outlook 2018

Abstract

This chapter introduces the 2018 edition of the Science, Technology and Innovation (STI) Outlook, distilling the main trends and policy issues from across the chapters into a few key highlights. It is organised into five main sections, starting with the drivers of change disrupting research and innovation and STI policy. Subsequent sections explore their impacts on innovation processes and scientific practices, and raise the question of how STI policy and governance practices can adapt to opportunities and challenges in a fast-changing context.

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

Introduction

Developments in science, technology and innovation (STI) are major drivers of change in modern societies. They are themselves subject to various influences – including a range of societal, economic and technological factors – that shape their activities and outcomes. Public policy is another important influence on STI, because of its funding and regulatory functions. Like its subject matter, STI policy is also subject to multiple influences on its agenda, design and implementation.

Four major trends influencing the direction and design of STI policy stand out. First, support programmes for public research and innovation face growing demand to demonstrate economic and societal relevance and impact. In particular, STI policy increasingly focuses on “challenges”, as governments seek to redirect technological change from existing trajectories towards more economically, socially and environmentally beneficial technologies, and to spur complementary private STI investments. This shift has given impetus to a new era of “mission-oriented” STI policy, with governments looking to work more closely with the business sector and civil society to steer the direction of science and technology towards specific goals.

Second, digitalisation is transforming science and innovation processes. Data have become a main input to innovative activities, and many innovations occur in software or data. Aspects of innovation are also accelerating as digital technologies shorten the time needed to perform some tasks. All areas of research are also becoming data-intensive, increasingly relying upon and generating big data. These changes have great potential to improve the productivity of innovation and science, but they require an adaptation of STI policies.

Third, many OECD governments have limited fiscal room for manoeuvre as they seek to reduce their debt burdens. As the latest data available show, current trends in government research and development (R&D) funding in the OECD area may not match the ambition and challenges inherent to mission-oriented policies. Under these conditions, it might be difficult for governments to make the investments in research and innovation activities needed to steer the direction of science and technology.

Fourth, governments can benefit from embracing digital technologies to design, implement and monitor STI policies. Digitalisation is already having a significant impact on the evidence base for STI policy and governance. The growing use of digital tools in research and innovation processes leaves more “digital traces”, i.e. digital data that can be used to produce indicators and analysis. Exploiting these traces will provide governments with more granular and timely data, to inform and improve science and innovation policies. Digitalisation can also help meet policymakers’ need to demonstrate the relationships between science and innovation expenditures and real-world outcomes.

This chapter introduces the Science, Technology and Innovation Outlook 2018, distilling the main trends and policy issues highlighted in the report. As such, it remains largely within the limits set by the chapters’ content, and does not aim to survey all of the main trends and issues affecting STI and STI policy today.

What are the economic, societal and technological drivers of STI policy changes?

Combining policy action to address rising economic and societal challenges

The 2016 edition of the STI Outlook described several megatrends that are expected to have a strong impact on research and innovation systems over the next 10-15 years and beyond (Box 1.1). These megatrends are quite slow-moving, which means they remain useful points of reference for thinking about economic, societal and political challenges that STI and STI policy will have to contend with (OECD, 2016). Many are addressed by the Sustainable Development Goals (SDGs) and articulated as “grand societal challenges” that increasingly shape STI policy agendas. At the same time, as these megatrends play out, the variety and degree of uncertainty they generate has unleashed reactionary forces in some countries that challenge much of the post-Second World War economic, political and social consensus.

Box 1.1. Selected megatrends affecting STI

Demography: The world population will continue to grow and is expected to nudge the 10 billion mark by the middle of the 21^st century. Africa will account for more than half of this growth, with a significant increase in the number of the continent’s young people. In other parts of the world – including in many developing countries – populations will significantly age: the number of people over the age of 80 will account for around 10% of the world’s population by 2050, up from 4% in 2010. With a declining share of the population in work, ageing countries will face an uphill battle to maintain their living standards. International migration from countries with younger populations could offset this decline, although it will likely meet resistance. Technologies that enhance physical and cognitive capacities could allow older people to work longer, while growing automation could reduce the demand for labour. Driven by this demographic increase and by the growing number of people living in large cities, the global population will be increasingly urban, with 90% of this growth occurring in Asia and Africa.

Natural resources and energy: The growing population, coupled with economic growth and climate change, will place considerable burdens on natural resources. Severe water stress is likely in many parts of the world, while food insecurity will persist in many predominantly poor regions, exacerbated by climate change. Energy consumption will also rise sharply, contributing to further climate change, absent significant uptake of renewables. Global biodiversity will come under increasing threat, especially in densely populated poorer countries.

Climate change and environment: Mitigating the considerable extent and impacts of climate change will require setting ambitious targets reducing greenhouse gas emissions. The latest report from the Intergovernmental Panel on Climate Change (IPCC, 2018) points out that several critical climate-change impacts could be avoided by limiting global warming to 1.5 degrees Celsius (ºC) compared to 2ºC, but that this will require rapid, far-reaching and unprecedented changes throughout society. Ambitious targets for waste recycling are also needed, implying the need for a major shift towards a low-carbon “circular economy” by mid-century. This shift will affect all parts of the economy and society, and will be enabled by technological innovation and adoption in both developed and developing economies.

Globalisation: The world economy’s centre of gravity will continue to shift eastward and southward, and new players – including governments, certain non-state actors (such as multinational enterprises and non-governmental organisations), and newly emerging megacities – will wield more power. Many of these shifts in power and influence are driven and facilitated by globalisation, which operates through cross-border flows of goods, services, investment, people and ideas, and is enabled by widespread adoption of digital technologies. However, globalisation will inevitably face counter-currents and crosswinds, such as geopolitical instability, possible armed conflict and new barriers to trade stemming from increased protectionism.

Health, inequality and well-being: The treatment of the infectious diseases that affect the developing world disproportionately is being compromised by growing antibacterial resistance. Non-communicable and neurological diseases are projected to increase sharply, in line with demographic ageing and the global spread of unhealthy lifestyles. Technological advances in DNA sequencing, omics technologies, synthetic biology and gene editing have given researchers new tools to decipher and treat chronic non-communicable diseases. Inequalities and poverty remain a concern in many developed countries, although global poverty continues to decline.

Source: Adapted from OECD (2016), OECD Science, Technology and Innovation Outlook 2016, https://doi.org/10.1787/sti_in_outlook-2016-en.

The necessary investments to address these challenges will likely be made in a difficult economic context. According to a recent long-term baseline scenario prepared by the OECD, annual global growth is estimated to slow from 3.5% currently to 2% in 2060 (Guillemette and Turner, 2018). Productivity growth has fallen over the past two decades, especially since the 2008 global financial crisis. This trend, combined with low or declining multi-factor productivity growth in several countries and sectors, has raised concerns about the ability of research and innovation activities to support economic growth and social well-being. Scholars continue to debate the reasons for the slowdown. Some point to slower rates of innovation, which is the root of productivity. Others point to the historical time lag between innovation and its impacts on productivity. They argue that the productivity crisis will end as businesses learn, relevant structural reforms are implemented, complementary investments are made, and recent innovations are broadly adopted and adapted beyond lead innovators. Another potential explanation for some part of the productivity slowdown is mismeasurement of the increasingly digital economy.

Solving increasingly pressing societal challenges at a time when financial resources in several OECD member countries are limited, and the growth engine seems to be stalling, will be difficult. It will require combined actions to solve economic, societal and environmental challenges. With the right policies and incentives in place – notably strong fiscal and structural reform, combined with coherent climate policy – governments can generate growth that will significantly reduce the risks of climate change, while also providing near-term economic, employment and health benefits. For instance, it is estimated that a climate-compatible policy package, including enhanced incentives for innovation, could increase long-run gross domestic product (GDP) by up to 2.8% on average across the Group of Twenty (G20) in 2050 (OECD, 2017a).

 

Societal and environmental challenges will need to be addressed in a difficult economic context, characterised by low global growth and productivity. Combined policy actions could help solve economic, societal and environmental challenges. Fuelled by innovation, an effective policy to combat climate change could also generate significant growth and new jobs, and enhance well-being.

New emerging technologies hold great potential

New emerging technologies can help address many grand societal challenges. Building on earlier OECD work (OECD, 2017b), Chapter 2 on artificial intelligence (AI) and the technologies of the Next Production Revolution provides many examples of emerging technologies with wide-ranging future applications. For example, gene editing could revolutionise today’s medical therapies; nanomaterials and bio-batteries could provide new clean-energy solutions; and AI could become the “primary-drug discovery tool” over the next decade. In the medium term, some technologies now at the demonstration stage could have significant impacts. For instance, new generations of bio-refineries, which transform biomass-waste products into marketable products and energy, have the potential to substantially reduce greenhouse gas emissions.

 

If well managed and used in conjunction with social innovation and policy reforms, scientific and technological advances have the potential to significantly alleviate many grand societal challenges.

Some of these technologies are already being applied. For example, AI, enabled by ongoing improvements in computer hardware, the widespread availability of large datasets and improved software, has growing applications in various areas of production, from semi-conductors and pharmaceuticals, to more traditional “bricks-and-mortar” industries like mining and construction. However, as Chapter 2 shows, the diffusion of AI and other advanced production technologies is by some measures quite slow, and policies that facilitate diffusion could benefit productivity growth.

Blockchain technology has recently attracted much attention. Together with robotics (e.g. using software robots for process automation) and AI (e.g. for detecting data anomalies and identifying process vulnerabilities), blockchain could significantly change key financial services, from financial transactions to automated contractual agreements. Blockchain technology was first applied in cryptocurrency markets. However, many other applications (e.g. remittances, inter-bank transfers and securities trading) are now emerging in the financial sector (OECD, 2017c) and, as Chapter 2 shows, blockchain is beginning to play roles in production as well. Furthermore, technological convergence – for instance, the combination of technologies such as the Internet of Things (IoT), blockchain, AI and advanced robotics – may open new production frontiers.

How are technological and societal change transforming innovation processes?

The very characteristics of the innovation process are changing as a result of technological opportunities (particularly stemming from the digital transformation of the economy), as well as societal pressures and growing aspirations for more inclusiveness and openness. These changes are unfolding in a more favourable business environment: firms have resumed their R&D investment since the financial crisis, fuelled by restored profitability and the increasingly generous R&D fiscal incentives offered by many governments.

Digitalisation is creating new opportunities for innovation and knowledge exchange

Chapter 3, on innovation policies in the digital age, analyses the impacts of digitalisation on innovation processes. Key phases of the innovation cycle are becoming faster and cheaper. The costs of searching, verifying, manipulating and communicating information and knowledge, as well as the costs of launching innovative goods and services in the market, are falling. Aspects of innovation are also speeding up as competition increases and digital technologies allow some tasks to be performed more quickly, for instance, in design and testing. The growing availability of digital data on customers’ needs, and the ability to experiment more easily with data on different customer groups, also helps streamline product and process innovation. As digital technology significantly lowers the cost of versioning, products can be differentiated (and even personalised) much further. As a result, the product cycle can be accelerated, changing the speed of market competition.

Data have become a major input to innovation: basic data on the characteristics of materials or the environment, or on customer demand, can be used to identify a product’s optimal features and create “digital twins” of machinery and physical goods, allowing deeper forms of process optimisation. Access to data has become a key parameter in business strategies: companies that control valuable and unique data have a competitive advantage over others. Contrary to physical inputs, data can be re-used and shared, creating new opportunities for collaboration between businesses.

Innovation has also become more collaborative, thanks to both improved conditions on the supply side (data sharing) and stronger demand for collaboration, stemming from increased interdisciplinarity and engagement with a variety of stakeholders. Collaboration can take several forms, such as data sharing, open innovation, digital platforms, and mergers and acquisitions. Interactions along global value chains, which have become increasingly important since the early 2000s, are also affected by digital technologies, with important consequences for the redistribution of high value-added activities among countries (e.g. reshoring of highly automated activities; see De Backer and Flaig, 2017).

The digital transformation has also supported the emergence of new forms of policy support for knowledge transfer. For example, online platforms, networks and communities have emerged as new spaces for knowledge transfer, helping to match supply and demand for technology. They connect firms with global networks of public research centres, individual scientists and freelancers that can help solve specific technological problems. Enhanced options for electronic exchanges have also led to the creation of new models of “off-campus” technology transfer offices (TTOs), such as TTO alliances at the regional, national or sectoral level. These typically result from co-operation between several universities and public research institutes (PRIs), as in Germany (the regional patent agencies) and France (the technology transfer acceleration companies), for example. Pooling specific resources and services (e.g. patent databases and services, marketing and communication activities, and training and experts) often improves efficiency and the quality of services provided by TTOs. Given the broad variety and distribution of these developments, policymakers can play a useful role in promoting integration, co-operation and interoperability between the patchwork of existing and emerging initiatives (OECD, 2018a).

New policy and business practices for inclusive innovation are emerging

Innovation delivers far more than new or improved products and services that provide companies with a competitive edge and contribute to economic growth. Innovation can also be “inclusive”, responding to the needs of a broader array of stakeholders. First, innovation can contribute to new or improved products and services for those at a social disadvantage. For example, innovation can provide lower-income groups with greater access to services such as long-distance calling, “e-learning”’ and “e-government”. The effect of technical progress on prices can also contribute to social inclusion: some information and communication technology (ICT) products, such as laptops and smartphones, have become increasingly affordable and available to a higher number of people. Second, the process of innovation itself can become more inclusive, as previously underrepresented individuals and social groups can now participate in it more easily (OECD, 2017d). Chapter 10 on technology governance presents some emerging business-innovation practices that are more open, co-creative and responsive to social needs. These practices sometimes offer opportunities for individuals and small groups to engage in digital production in dedicated small-scale sites, e.g. maker spaces, living labs and fab-labs. These local workshops are more accessible to potential “non-traditional” innovators – especially young innovators and independent inventors – and are often based on collaboration with universities and local authorities. Innovation can thus become a factor of social inclusion as participating groups develop new skills and broaden their range of opportunities.

 

Innovation can be inclusive. Several open, co-creative and socially responsive practices are emerging. Dedicated sites, such as maker spaces, living labs and fab-labs, are now found in most countries and support the activities of potential “non-traditional” innovators.

Established firms can also engage in inclusive innovation practices. Chapter 10 on technology governance identifies some more inclusive and open practices (e.g. design ethics) that firms are using at early stages in the innovation cycle. Although such practices are recent and still emerging, they could be powerful tools for translating and integrating core social values, safeguards and goals into technology development. In the field of nanotechnology, for instance, standardisation is seen not only as a means of facilitating commerce through interoperability, but also of promoting health and safety. For example, it can embed knowledge of potentially adverse effects in the design of nanomaterials and nanoproducts. In many initiatives, the value added lies as much in the result – i.e. “ethical” technologies and products – as in the process itself. In addition to their usual technical work, some standardisation committees (like the Institute of Electrical and Electronics Engineers) also operate as fora for public discussion on issues such as AI.

Government support for business R&D is shifting

Business firms have an essential role in developing, diffusing and using the new wave of technologies. This requires major investment in R&D, as well as in other complementary assets and intangibles in a wide variety of fields. The analysis of business R&D expenditures (BERD) shows that firms have taken up this challenge. BERD has picked up in many countries since the financial crisis and is almost back to its pre-crisis growth trend, both in volume and relative to GDP. This increase is driven by growth in aggregate demand and firms’ restored profitability (Figure 1.1, Panel a). It is also driven by relatively new actors in the R&D field – mainly large firms in digital industries, which are investing massively in AI and other Next Production Revolution technologies.

Although the bulk of business R&D is financed by companies, public support helps incentivise these activities and focus them on certain public-policy priorities. Global trends in public support for business R&D are difficult to interpret, as policy approaches differ markedly across countries (Figure 1.1, Panel b). However, the share of BERD that is funded by government through direct support (such as grants) has dropped in all countries since the financial crisis, from 14.1% (2009) to 6.8% (2016) in the United States, and from 7.3% (2010) to 6.3% (2015) in the European Union.

Figure 1.1. Trends in business R&D financed by businesses and government

Note: The black line in panel b is BERD financed by government in OECD countries, less the United States.

Source: Calculations based on OECD (2018d), "Research and Development Statistics: Government budget appropriations or outlays for RD (Edition 2017)", OECD Science, Technology and R&D Statistics (database), https://doi.org/10.1787/e724dc33-en (accessed on 26 September 2018)

Statlink

However, this decrease in direct support for business R&D has been amply compensated by an increase in indirect support through tax incentives over 2006-14 (OECD, 2016, 2017e). When considering total (direct and indirect) government support for business R&D, a majority of countries (i.e. 29 out of the 41 countries for which data are available) increased their support to business R&D, relative to GDP, over 2006-15. This increase is particularly significant in countries where tax incentives account for a large share of total government support (Figure 1.2). It is often related to the reform of indirect support schemes for business R&D, to make them more available, accessible and generous; 12 OECD countries also introduced such schemes over 2000-15 (OECD, 2018b). Thus, the share of tax relief in total government support for business R&D in the OECD area increased on average from 36% to 46% between 2006 and 2015 (OECD, 2018c).

Figure 1.2. Direct government funding and tax support for business R&D, 2015 and 2006

Source: OECD (2017e), OECD Science, Technology and Industry Scoreboard 2017: The digital transformation, https://doi.org/10.1787/9789264268821-en.

Statlink

Even though R&D tax incentives are considered more cost-efficient and easier to operate than subsidies and grants, they are exclusively allocated to business R&D that reflects market needs. By design, but also owing to specific regulations (e.g. the European Union’s Community State Aid rules for R&D and innovation), it is difficult to designate specific fields of research or sectors that would most benefit from these indirect incentives. As a demand-driven policy tool, indirect public funding also offers little margin for governments to influence the amounts allocated, apart from (for example) “capping” the allocated credits. Thus, this shift in the policy mix raises the issue of governments’ capacity to influence the direction of private R&D, at a time when achieving societal and environmental goals requires more – and more focused – innovation.

How is science evolving to become more open, automated and gender-friendly?

The 2016 edition of the STI Outlook provided a high-level overview of the main trends and issues set to shape science systems over the next 10-15 years (Box 1.2). While these are still valid, the opportunities and challenges they present continue to unfold, as do the policy responses. Some issues have gained importance in the last two years. These notably include the impact of digitalisation – which this edition of the STI Outlook covers extensively – and the “reproducibility crisis” in science, whereby a growing number of results in scientific publications are difficult or impossible to reproduce by other researchers. The accelerating rollout of open science also places greater emphasis on transparency in science: open-science principles stress open-access publication, open data sharing, and more open and inclusive participation in science itself. Another growing concern is how to support breakthrough research when faced with seemingly decreasing research productivity, and societal challenges of unprecedented scale and scope. It is difficult to demonstrate through data analysis or case studies that new ideas are becoming “harder to find” (Bloom et al., 2017; Jones, 2009). However, several research communities claim that competitive funding mechanisms disadvantage risky, potentially transformative and transdisciplinary research proposals in favour of applied, incremental and mono-disciplinary proposals.

Box 1.2. Key science systems trends and issues

The 2016 edition of the STI Outlook included a chapter on the future of science systems, featuring several key trends and issues that were expected to shape science systems over the next 10-15 years. These trends include:

• fiscal restraint and competing policy demands, placing pressure on government R&D spending

• the growing importance, in some research systems, of non-state funding for public research, including by philanthropists, charities and foundations

• the growing share of public research performed by emerging economies – particularly China, which is now second only to the United States in its overall public expenditure on R&D

• the re-orientation of public science agendas towards “grand societal challenges”, with a growing emphasis on the SDGs as a framework for agenda-setting

• the turn towards more challenge-driven public research, placing more emphasis on interdisciplinary research and the interfaces between basic and applied research

• emerging new arrangements for commercialising public R&D, including new TTO-type structures and the use of smarter IP strategies in public research-performing organisations

• growth in citizen science, including “do-it-yourself” science

• greater consideration of the ethical, legal and societal aspects of research, within a framework of “responsible research and innovation”

• emerging new opportunities from the growing digitalisation of science (e.g. regarding automation, big data and more open science), but also significant challenges (e.g. regarding data ownership, conflicting incentives for open science, the costs of maintaining data infrastructures and the availability of skills)

• the growing precariousness of research careers in hyper-competitive research environments, and its negative impacts on certain groups (particularly women)

• shifts in the ways of assessing research performance to reflect the emergence of non-traditional bibliometrics (“altmetrics”) and the greater use of public-value criteria to assess the contributions of research to societal challenges

• growing concerns about the “reproducibility crisis” in science

• the growing gap between scientific evidence and other forms of knowledge and opinion, complicated by the global, multi-dimensional, fast-evolving and complex nature of many grand societal challenges.

Source: OECD (2016), OECD Science, Technology and Innovation Outlook 2016, https://doi.org/10.1787/sti_in_outlook-2016-en.

Building upon the wide-ranging assessment of 2016, the 2018 edition of the STI Outlook examines three prominent topics in the current debates on research policy. The first topic is open science and enhanced access to research data, which has several potential benefits, but also faces significant challenges. The second topic is the impact of AI and automation on science, which has the potential to transform science practice over the next decade. The third topic is the long-standing under‑representation of women in certain areas of science; while many policy interventions seek to address the issue of gender in science, much remains to be done.

Enhanced access to research data has many benefits

All areas of research are becoming increasingly data-intensive, and big data are no longer the prerogative of experimental physics and astronomy. Chapter 6 highlights the expected benefits of enhanced access to data, i.e. new scientific breakthroughs, less duplication and better reproducibility of research results, improved trust in science and more innovation. These benefits, however, should be balanced against the costs, including the need to protect privacy and security, and prevent malevolent uses. Accordingly, “as open as possible, as closed as necessary” is gradually replacing the “open-by-default” mantra associated with the early days of the open-access movement.

Enhanced access to data poses several outstanding policy challenges. First, governments need to put in place systems and processes to ensure transparency and foster trust across the research community and wider society. For example, while privacy breaches cannot be avoided, the risks should be managed, and the procedures to do this should be clear and transparent. Second, implementation of the FAIR Guiding Principles1 for policy development and co-operation across communities depends on the development and adoption of a common technical framework (Wilkinson et al., 2016). Policymakers should therefore support bodies (such as the Research Data Alliance) that are building the social and technical infrastructure to enable open data sharing across national and disciplinary borders.

Third, appropriate recognitions and rewards need to be in place to encourage researchers to share data. Data activities should be embedded in evaluation systems to ensure that researchers who provide high-quality research data (including on negative results) are rewarded. Generalising data citation, so that it can be used to incentivise and reward data sharing, also requires new data-citation metrics.

Fourth, the substantial costs of data stewardship and provision entail a long-term financial commitment. Funding them requires understanding not only the business models and value propositions of specific data repositories, but also the research networks in which they are integrated. In some instances, it may make sense to centralise the management of data resources to obtain economies of scale across research systems.

Finally, the additional burden of curating and stewarding data to make them openly available for secondary use is a science-wide human-resource challenge, which will only be met through retraining existing personnel, and providing new education and training opportunities to researchers and professionals in research data-support roles. Data scientists are in high demand in industry, and academic research competes for the best talent – hence the urgent need to develop attractive career paths, to realise the value of enhanced access to public research data.

 

Government should help science cope with the challenge of open science. This includes ensuring transparency and trust across the research community and wider society; enabling the sharing of data across national and disciplinary boundaries; and ensuring that recognition and rewards are in place to encourage researchers to share data.

Automation could transform scientific practice

AI and machine learning have the potential to increase the productivity of science, enable novel forms of discovery and enhance reproducibility. AI in science has already predicted the behaviour of chaotic systems, tackled complex computational problems in genetics, improved the quality of astronomical imaging, and helped discover rules of chemical synthesis. Since AI systems have very different strengths and weaknesses compared to human scientists, they are expected to augment human abilities in science. Chapter 5 outlines three key technological developments driving the recent rise of AI: improved computer hardware, increased availability of data and improved AI software. Examples are rapidly accumulating of AI being applied across the entire span of scientific enquiry.

Broadening the use of AI in science faces several challenges. First, despite the impressive performance of AI in many areas, the need still exists to further develop AI methods that perform well in constrained and well-structured problem spaces so as to be able to apply these to scientific domains where data are noisy and corrupted and processes are only partially observed. This need exists in climate science, for instance, where the number of variables involved is vast, uncertainty exists on which feedback loops are important, and accurate measurement – although improving – is still a challenge. Creating approaches that work across all data scales – from data-sparse environments to data-rich contexts – will be key. Second, discussions of AI commonly cite the lack of transparency in machine learning-based decision-making as a source of possible concern. As Chapter 5 points out, questions of intelligibility are not confined to machine learning (only a few specialists understand the proofs involved in some leading areas of mathematics, for example). Some existing techniques provide audit trails of machine learning and can help explain its results. But the question of intelligibility is likely to become more salient as AI techniques are used more widely. Third, education and training is a key policy issue. Too few students are trained to understand the fundamental role of logic in AI; bridging this gap will require changes in curricula. Finally, the computational resources for leading-edge AI research are enormous, and can be expensive. The largest computer resources and the largest number of excellent AI researchers are found in the business sector, not in public science.

 

Several challenges hinder the widespread use of AI in science: the need to transform AI methods to operate in challenging data conditions; concerns regarding transparency in machine learning-based decision-making; the need for more, and more tailored, education and training in AI; and the cost of computational resources for leading-edge AI research.

Removing gender barriers in science requires more joined-up policy

Chapter 7 on gender in the changing context for STI reviews the key issues affecting gender equity in science at different life stages. Gender stereotypes influence educational choices and career expectations even in early childhood. In undergraduate and graduate education, women and men are also unevenly distributed across academic courses, with women significantly less represented in certain science, technology, engineering and mathematics (STEM) fields (particularly engineering, ICT, physics, mathematics and statistics). At the doctoral level, on the other hand, the share of women in certain STEM fields has increased over time, and the “leaky pipeline” between graduate and postgraduate education and training is no longer a major challenge (Miller and Wai, 2015). In research careers, early-stage researchers often hold precarious positions in very competitive environments. Hyper-competition, and its reinforcement of assertive stereotypes, serves as an exclusionary mechanism for those who cannot or will not compete continually. The choice to enter this competition often coincides with “the rush hour of life”, i.e. the establishment of partnerships and families, thereby reinforcing gender imbalances.

The changing context for STI increases the need for diversity. While social justice and fairness are important issues in themselves, increasing evidence shows that diversity improves the quality of research outcomes and their relevance for society (Smith-Doerr et al., 2017). Diversity and inclusiveness in STI are a prerequisite for producing the types of knowledge and innovations required to respond effectively to all the SDGs.

Against this backdrop, Chapter 7 lays out a future vision for a more diverse and productive scientific enterprise that recognises and rewards the equivalent and distinct contributions of both men and women. Most countries have adopted this objective, with many national plans identifying gender equity as a strategic priority. The 2017 edition of the EC/OECD STI policy survey shows that this priority has been translated into many specific policy initiatives related to gender in STI. However, the overall policy picture today also points to a fragmented approach, characterised by multiple institutions acting independently and limited co-ordination between education, science and innovation actors. Little systematic evaluation takes place of the effectiveness and long-term impact of the many interventions under way. Engaging in strategic thinking and targeted interventions to create positive feedback loops and strengthen the position of women within STI systems will require co-ordinated actions across actors at multiple levels.

 

Most countries cite gender diversity as one of the key objectives of their national STI plans. However, policy initiatives remain fragmented. A more strategic and systemic long-term approach is necessary.

How is STI policy responding to societal and technological disruptions?

As disruptive technologies create new challenges and opportunities, the terms of reference for STI policy making are changing. Meeting societal challenges has become a prominent goal, and mission-oriented policies to do this, with defined goals and within defined timeframes, are increasingly popular. However, governments’ capacity to engage in such directive policies and significantly affect the major outcomes – particularly since the share of R&D in government spending has declined overall in OECD member countries – is questionable.

Societal challenges: From shaping the STI agenda to influencing specific policy actions

As revealed by a recent survey reported in Chapter 9, on the governance of public research policy, most OECD countries have STI strategies explicitly referencing societal challenges. Out of the 35 countries surveyed, 33 (94%) have a national STI strategy or plan in place. Meeting major societal challenges is an objective in most of these strategies (30 (90%) of 33 strategies). Key priority themes include sustainable growth, health improvements and efficient transportation systems. Strategies often refer to the SDGs, which have become an important political framework globally. However, as shown in Chapter 4 on STI policies for the SDGs, addressing societal challenges is rarely the main rationale for STI policy initiatives, although many competitive-funding schemes include societal impacts as selection criteria. References to STI in the SDGs are often more implicit than explicit. Greater effort outside of STI policy arenas is needed to demonstrate the role of research and innovation in helping to meet the SDGs. This will require a closer alignment of existing STI governance structures (e.g. policy advice, steering and funding, co-ordination, evaluation and monitoring) with the emerging “global governance framework” for the SDGs.

 

There is a lack of explicit reference to STI in the SDGs, and too little reference to the SDGs in STI. STI governance structures should be more closely aligned with the emerging “global governance framework” for the SDGs.

The magnitude and transnational scope of global challenges and the size of investment needed to address them, demands international co-ordination and co-operation of research efforts. International co-operation in STI provides parties with access to knowledge and expertise and enables cost sharing while avoiding duplication of research efforts. International co-operation among scientists has never been higher nor more diversified as shown by data on co-authored publications and co-patenting. At the same time, however, an erosion of multilateralism in other policy areas threatens international cooperation in STI. The challenge for STI policymakers is to demonstrate the benefits of international co-operation more forcefully in terms of economic, societal and environmental impacts. International STI co-operation for global challenges will also require mechanisms to ensure equitable sharing of the burden of global research efforts as well as the benefits.

Toward a new type of strategic steering to cope with economic and societal challenges

There are growing calls to support economic growth and address societal challenges through strategic steering of STI. As Chapter 4 on STI policies for the SDGs points out, reframing STI policy is not straightforward, and pleas to transform policy frameworks have yet to outline clear pathways for policymakers and propose new levers for policy. At best, they have suggested incremental reformulation of traditional supply and demand-side instruments, by instilling considerations of sustainability and directionality.

Against this backdrop, new mission-oriented programmes have been proposed, for example in the context of the preparatory discussions for the European Union’s “Horizon Europe” plan. Mission-oriented programmes are large-scale interventions aiming to achieve a set mission (goal or solution) within a well-defined timeframe, with an important R&D component. Missions are more concrete than broad grand challenges, because they have clear time-bound targets. Compared to previous mission-oriented policies, the new missions focus more clearly on the demand side and the diffusion of innovations, seek to be coherent with other policy fields, and recognise the roles of both incremental and systemic innovations. They are intended as “systemic” public policies that draw on frontier knowledge to attain specific, often very ambitious, goals (Mazzucato, 2018). The terms of reference of these new mission-oriented programmes are still under development; they include melding the entrepreneurial power of bottom-up projects and the “purposive” top-down steering necessary for transformative innovation.

There exist multiple examples of failed mission-oriented policies. Lessons from these experiences warrant caution and attention to the design and evaluation of mission-oriented approaches. While certain examples of bold programmes can be found in history (notably in the space and defence industries), applying their lessons to another context and/or era will call for different policy and governance arrangements. A major difference is that in many previous missions, the government was the main or sole purchaser of the resulting technological developments. Government labs were also often the main performers of R&D. Today, the private sector performs most R&D in many OECD countries. Moreover, undertaking missions dedicated to grand societal challenges will require significant levels of funding and specific co-ordination mechanisms, involving companies and civil society actors. This means that governments need to favour public-private partnerships, where risks and rewards can be shared. Governments are using deliberative processes to better align innovation strategies and societal priorities. Nevertheless, questions remain over governments’ capacity to add directionality to STI processes, given their limited fiscal room for manoeuvre and their existing sets of skills and capabilities.

 

New mission-oriented programmes could mobilise science and innovation to address societal and economic challenges. However, their governance and design arrangements have yet to be developed and tested.

Supporting the development and uptake of emerging technologies requires a mix of old and new types of policy interventions

Developing and using effectively and ethically new technologies involves various changes to STI policy making and governance. Some are technology-specific, while others are more cross-cutting. Several new digital technologies call for new types of intervention that require further experimentation and learning. Various chapters, notably Chapter 2 on AI and the technologies of the Next Production Revolution, Chapter 3 on digital innovation and Chapter 6 on enhanced access to data discuss some new tasks for governments. For example, governments can help support the development and sharing of data as part of open-data initiatives. They can act as catalysts and honest brokers in data partnerships, e.g. by co-ordinating and stewarding data-sharing agreements. Although such efforts are generally undertaken at national level, several international and multilateral initiatives have emerged to foster open access to STI data.

At the same time, since the traditional rationales for public intervention remain valid for government support of emerging technologies, STI policy and governance also need to deliver existing policies more effectively. For example, the technologies of the Next Production Revolution (including microelectronics, synthetic biology, new materials and nanotechnology) result from advances in scientific knowledge and instrumentation. Government support is essential to promote basic research, and to provide incentives and appropriate conditions for effective science-industry relationships. Even AI, research on which is led today by large private companies, rests on decades of public research that provided the foundations for today’s developments.

Diffusion-oriented policies are also crucial. For complex systems, such as biorefineries, public-private partnerships around demonstrators can help to resolve technical and economic questions about production before the necessary large investments are made. Governments also have roles to play to help small and medium size enterprises understand and eventually adopt emerging technologies. Further downstream, the certification of technologies, such as 3D printing, will support their diffusion by controlling for possible negative impacts, e.g. related to the risk of environment damage.

Numerous specific challenges may hinder the necessary policy changes. Since the Next Production Revolution has implications for a wide range of fields (including digital infrastructure, skills and intellectual property rights) that were previously not closely coordinated or connected in government, it could accentuate co-ordination problems already apparent in many countries. Governments also often lack knowledge and skills in many areas of complex and fast-evolving new technology. Supporting the transition to Industry 4.0 challenges governments to act with greater foresight and technical understanding across multiple policy domains. Accelerated innovation also raises challenges in providing targeted support, as targets may change so rapidly that traditional instruments could become irrelevant. Governments need to adapt: adopting broader targets, moving targets and flexible management are possible avenues.

 

Several emerging technologies arise in a wide range of fields not previously closely connected in government, creating co-ordination problems. Many governments also lack knowledge and skills relevant to complex, fast-evolving new technologies.

Reaping the benefits of emerging technologies to ensure economic and social progress requires substantial and effective investment in research and innovation

Although the quality and type of research and innovation are as important as the absolute funding amounts allocated, all the policy initiatives outlined above require financial resources. However, whether current trends in public R&D efforts are commensurate with the current and future challenges needing to be addressed is an open question. Government budget allocations for R&D (GBARD) typically rose before the crisis. A few years after the crisis, however, once the additional spending related to stimulus packages and recovery plans had been exhausted, GBARD decreased or flattened in all Group of Seven countries (G7), except Germany. Given that these countries had the largest R&D budgets, GBARD has declined overall in the OECD area (Figure 1.3).

Figure 1.3. Government budget allocations for civil R&D, 2000-08 and 2008-17

Note: GBARD less defence.

Source: Calculations based on OECD (2018d), "Research and Development Statistics: Government budget appropriations or outlays for RD", OECD Science, Technology and R&D Statistics (database), https://doi.org/10.1787/data-00194-en (accessed on 14 September 2018).

Statlink

 

The current trend in public R&D spending may not match the current and future challenges that science and innovation must address. Since 2010, government R&D expenditures in the OECD as a whole and in almost all G7 countries have stagnated or decreased not only in absolute amounts and relative to GDP, but also as a share of total government expenditure.

Comparing the evolution of public budgets for R&D with the budgets for all policy domains combined sheds further light on public funding dynamics (Figure 1.4). A positive correlation exists between the evolution of overall government budgets, and the evolution of the budget for R&D. It is reasonable to assume that the overall budget is one driver of the R&D budget, as governments consider their overall financial position before allocating funds across budget lines. Hence, the slowdown in R&D spending might be partly explained by overall budget restrictions following spending on the recovery packages of 2009 and plans to moderate or reduce public debt. Accordingly, all countries where GBARD increased also experienced an increase in their overall government expenditure (top-right corner). In several countries, public R&D budgets have also increased more rapidly than the overall public budget, thereby increasing the share of R&D in government spending (top right corner, above the diagonal line in the graph). However, since six of the G7 countries experienced an opposite trend, government R&D funding has decreased as a share of total government expenditures for the OECD area as a whole. As discussed in Chapter 8 on new public research-funding approaches and instruments, this might suggest that the policy importance of research and innovation has shifted downwards in many countries. More anecdotally, it echoes some policy officials’ frustrations – especially in finance ministries and centres of government – over the absence of sufficiently tangible innovation results stemming from the significant recovery plans implemented in the wake of the financial crisis.

Figure 1.4. Average annual growth of total government budgets and GBARD, 2009-16

Source: Calculations based on OECD (2018e), "General Government Accounts, SNA 2008 (or SNA 1993): Main aggregates", OECD National Accounts Statistics (database), https://doi.org/10.1787/data-00020-en (accessed on 8 October 2018); and OECD (2018f), "Total GBARD (Government budget allocations for R&D) at current prices and PPP", in Main Science and Technology Indicators, Vol. 2018/2, https://doi.org/10.1787/msti-v2018-2-table57-en.

Statlink

The share of government in total funding of R&D decreased by 4 percentage points (from 31% to 27%) in the OECD area between 2009 and 2016 (Figure 1.5); it only increased in five countries. Hence, the weight of government in total R&D funding has dropped, given growth in business expenditure on R&D has recovered. As previously discussed (Figure 1.2), adding R&D tax credits to public R&D budgets modifies the overall picture, since the tax credits increased significantly during this period. However, tax credits do not enhance government’s capacity to influence the direction of R&D, as they are direction-neutral by design. Accordingly, the reduced share of government in R&D funding could lead to less government influence on the overall direction of science and innovation.

 

The shift of the policy mix towards R&D tax incentives decreases governments’ capacity to influence the direction of private R&D towards socially desirable goals, at a time when the need for a more strategic orientation of research and innovation is becoming more pressing.

Business R&D has picked up in several countries in recent years and can therefore compensate somewhat for lower public spending. However, most firms focus on applied research and experimental development. Funding for basic research – without which many of the new developments linked to the digital revolution would not have happened – may be particularly at risk in the coming years.

Figure 1.5. Change in the share of government in the direct funding of gross domestic expenditure on R&D, 2009-16 (or latest year available)

Source: OECD (2018g), "Main Science and Technology Indicators", OECD Science, Technology and R&D Statistics (database), https://doi.org/10.1787/data-00182-en (accessed on 2 October 2018).

Statlink

Faced with government austerity measures, politicians and senior public-sector leaders in many countries are increasingly demanding hard evidence on the outcomes of research funding; they want to know what works, and what does not. Science and innovation spending is no longer exempt from pressures to provide quantitative evidence of impact. Against this backdrop, policymakers need to shift more attention to supporting monitoring systems, evaluation frameworks and data infrastructures (Chapter 12). As reported in Chapter 9 on the governance of public research policy, 19 of the 34 OECD countries surveyed have independent specialised agencies in charge of evaluating and monitoring the performance of higher education institutions (HEIs) and PRIs.

As shown in Chapter 8 on new public research-funding approaches and instruments, the growing demand for evidence of economic and societal impacts, in addition to scientific excellence, also affects the traditional modes of allocating government funds to public research institutions. Funding instruments have become more complex to respond to the growing number of economic and social objectives to be met by science and innovation. Although the range of options available to policymakers has expanded beyond traditional institutional “block” funding and competitive “project funding”, these instruments’ growing complexity and diversity creates new challenges (e.g. related to co-ordination and evaluation).

How is STI governance adapting to a fast-changing context?

Confronted with a more rapidly changing and varied research and innovation landscape, governments need to become more agile, more responsive, more open to stakeholder participation and better informed. Governments are experimenting with new approaches to policy design and delivery. They will also benefit from embracing digital technologies when designing, implementing and monitoring STI policies. A new generation of digital tools can produce more granular and timely data to support policy formulation and design. By linking different datasets, these tools can transform the evidence base for STI policy, and help demonstrate the relationships between science and innovation expenditures and real-world outcomes.

New modes of STI governance are emerging

Technological and social changes mean that the way governments work and interact with their policy subjects and partners is shifting. For example, new technologies like AI and gene editing can alter – and even disrupt – society and the economy in unpredictable ways. Preventing, correcting or mitigating the negative effects of technology has also become more important – yet more difficult – as technology itself has become more complex and pervasive. The fast pace of technological change means that policymakers struggle to exert oversight regarding emerging technologies. Traditional “end-of-pipe” regulatory instruments (such as risk assessment) are insufficient under conditions of uncertainty: they often fail to anticipate or address the long-term implications of emerging technologies, and they can be inflexible, inadequate and even stifling for innovation.

The uncertainty and risks created by rapid technological change cannot be borne and directed by the private sector alone: governments must take an active role. Chapter 11 on new approaches in policy design and experimentation argues that governments must evolve to better anticipate, adapt to and mitigate these change processes, as part of their STI policy portfolios and policy-making practices. However, policymakers face the extremely challenging task of balancing the need to maintain stability and confidence in the public system, while rapidly adapting to a new environment and new demands. Yet they must adapt, or governments risk becoming increasingly irrelevant, dysfunctional and disconnected. Despite their potential benefits, many emerging policy approaches – such as design thinking, collective intelligence, behavioural insights, policy experimentation and anticipatory governance – have yet to be widely adopted in STI policy.

 

Governments face a crucial trade-off between relevance and stability: they must evolve to better anticipate, adapt to and mitigate rapid technological and social changes, while maintaining stability and confidence in the public system.

New forms of technology governance are also required to allow policymakers to respond to technological change in real time. Technology governance is defined as the process of exercising political, economic and administrative authority over the development, diffusion and operation of technology. Chapter 10 argues that technology governance must move “upstream” and become an integral part of the innovation process itself, to steer emerging technologies towards better collective outcomes. This calls for more anticipatory and participatory modes of governance.

Anticipatory approaches can help explore, consult widely on and steer the consequences of innovation at an early stage. They can incorporate public values and concerns, mitigating potential public backlash against technology. They require new capacities within government that blend foresight, engagement and reflexivity to facilitate the acceptance of new technologies, while at the same time assessing, discussing and preparing for new technologies’ intended and unintended economic and societal effects. New policy tools, such as normative codes of conduct, test beds, regulatory sandboxes and real-time technology assessments can be useful. Chapter 3 on digital innovation and Chapter 2 on AI and the technologies of the Next Production Revolution also highlight the benefits of environments that facilitate learning, such as test beds and regulatory sandboxes, to help understand the regulatory implications and responses to emerging technologies. Participatory approaches can provide a wide range of stakeholders – including citizens – with effective opportunities to appraise and shape technology pathways.

These practices can help ensure that the goals, values and concerns of society are continuously enforced in emerging technologies as they unfold and that policymakers (and society) will not be taken by surprise. In doing so, they help shape technological designs and trajectories, without unduly constraining innovators.

Towards the next generation of STI data and indicators

Bringing together discussions and perspectives shared at the OECD Blue Sky Forum 2016, Chapter 14 on measurement and analysis in STI presents several key trends affecting the production and use of STI data and statistics. These include the increased connectivity of STI systems across national borders; accelerating digitalisation and its impacts on data availability and integrity; and pressures to demonstrate the impacts of public expenditures on STI in an era of government austerity. These drivers of change create new demands for STI data. They also question traditional forms of statistical definition and classification, which struggle to capture more fluid identities, attitudes and economic pathways. In addition, greater abundance of data place a premium on high quality, trustworthy sources, contributing to redefining the role of STI data experts and providers.

Given the increasingly globalised nature of science and innovation activities, the difficulties of national statistics in capturing the creation and circulation of knowledge and related financial flows across countries are a major concern. International collaboration is therefore required. While innovation systems do not change overnight, in times of fast-paced, disruptive change, timelier and more frequent data also become more critical. Timeliness is also essential for measuring processes that may be short-lived, such as entrepreneurship and business dynamics.

A key issue for STI policymakers is monitoring the link between science and innovation on the one hand, and the whole range of global sustainability concerns on the other, from poverty and hunger eradication, to equality and climate action. Those links are not easily traced, nor are they easily exposed through the sole use of indicators. The multidimensional nature of the SDGs implies that monitoring and measuring the overall role of science and innovation in meeting the SDGs will require accumulating findings from multiple sources. Chapter 4 on STI policies for delivering the SDGs also addresses this issue. It suggests that detailed administrative data and the ability to combine with other data could provide information on the role played by STI “input” commitments to the SDGs. However, obtaining and interpreting such data across countries presents some challenges.

Chapters 4 and 14 coincide in calling for policy frameworks, such as those being developed to monitor SDGs, to consider in their development an appropriate mix of instruments and disciplines for measurement and analysis. This should help address evidence needs and develop solutions that can be globally scaled up to achieve international comparability as well as greater synergies in highly interconnected STI systems. Defining and acting upon such needs requires a more strategic engagement between data producers and policy makers. Evidence from several countries and international initiatives suggest this is a feasible vision.

 

Monitoring the contribution of science and innovation to the global and multidimensional “grand challenges” remains difficult and will require new statistics and indicators. This effort should be linked to current work to develop indicators to measure overall progress on the SDGs.

The impact of digitalisation on the evidence base for STI policy and governance

Digitalisation is already having a significant impact on the evidence base for STI policy and governance. As more digital tools are used in research and innovation processes, they leave more “digital traces” that can be used for indicators and analysis. At a time when the cost of developing new data sources for responding to specific policy questions can be prohibitive, linking different existing data sources can provide insights that are impossible to obtain by working with the different data components separately.

Chapter 12 outlines the promises of digital science and innovation policy (DSIP), including: (i) streamlining burdensome administrative procedures, to deliver significant efficiency gains within ministries and agencies; (ii) providing more granular and timely data analysis to support STI policy, and improve the allocation of research and innovation funding; (iii) improving the timeliness of performance-monitoring data, to enable more agile short-term policy adjustments; (iv) detecting emerging patterns of change and stability in research, technology and industry, to support short-term forecasting of issues of policy concern; and (v) promoting inclusiveness in STI policy agenda-setting, by opening policy intelligence data to a broader range of stakeholders.

At the same time, policymakers’ expectations of DSIP infrastructures should avoid a naïve rationalism that understates the inherent complexity of policy making. DSIP systems can inform policy choices, but they cannot and should not provide a technical fix to what are ultimately political judgements, shaped by competing values and uncertainty. If open by design, DSIP systems could nevertheless be instrumental in embedding various social values in policy making by promoting inclusiveness in science and innovation agenda-setting, making it less technocratic and more democratic.

 

DSIP systems cannot provide a mere technical solution to policy making, which remains inherently complex and based on political judgements. But they could help embed various social values in policy decisions, by promoting inclusiveness in science and innovation policy agenda-setting.

Irrespective of the policy setting, an embedded, routine use of DSIP will depend for their adoption not just on digital technologies, but also on favourable social and administrative conditions. Organisations and individuals also need to have assurance that data about their funding, activities and results will be handled appropriately and protected when needed.

Chapter 13 on targeting entrepreneurship support on firms with high growth potential presents one possible application of big data and machine learning to a real-world STI policy problem (Box 1.3). It highlights some of the potential benefits, but also the limits of using these techniques in STI policy. Chapter 14 on measuring STI raises other concerns about using big data and machine learning, for example: the possibility that datasets contain defects and biases; difficulties in evaluating big-data techniques and analysis; and complexities in explaining these techniques to decision-makers and the public.

Box 1.3. Targeting entrepreneurship support on high-growth potential firms

Considering that only a tiny minority of new firms contribute to economic growth, some scholars have questioned the effectiveness of untargeted entrepreneurship policy, arguing that public resources should be concentrated on firms with the highest growth potential. This, in turn, poses the related question of whether it may be possible to identify high-potential firms ex ante. One difficulty in identifying successful entrants is the lack of detailed data on the characteristics of firms and entrepreneurs at the moment they create the company. As many firms are very small, limited public information is available from administrative sources. In this challenging context, policymakers can use big data and innovative predictive analytics (e.g. machine learning) to help target successful high-growing entrants.

There are important caveats in using such digital tools in this way. Significant unpredictability will remain about start-up success, as idiosyncratic and unobservable factors will always play an important role in rapidly changing markets. Periods of disruptive changes do not lend themselves well to policies aiming to pick the “best” firms for targeted support. Most innovations in turbulent times emerge through trial and error among various combinations of technological and social innovations. In such contexts, a subset of firms with higher growth potential are not ”revealed” to the world; their potential for growth emerges and increases through interactions with their environment, allowing faster learning and greater investment for some. Hence, direct and targeted policy interventions will always have to be complemented with horizontal reforms, to ensure an overall business environment conducive to entrepreneurship and experimentation. In practice, this means striking the right balance between targeting and promoting experimentation. However, evidence generated through big data and machine-learning techniques could influence this balance in the near future, pushing it more towards targeting (Chapter 13).

Private sector companies are increasingly contributing to the evidence base for STI policy, for example, as owners of bibliographic databases and providers of add-on services. The digitalisation of STI policy presents further opportunities for private-sector involvement. Although this presents several benefits, relying on the private sector for DSIP systems and components also creates potential risks for the public sector. For example, reliance on proprietary products and services may lead to discriminatory access to data, even if the data concern research activities funded by the public sector. Moreover, the public sector’s adoption of commercial standards for metrics may drive the emergence of private platforms exhibiting network effects that are difficult to challenge.

Conclusion

Taking various angles and approaches, the STI Outlook 2018 focuses on the policy changes needed to respond to the disruptions currently unfolding in technology, the economy, the environment and society. The 13 thematic chapters and this overarching introduction focus on many of the key policy questions. Taken together, they provide insights on the challenges at stake and a range of possible policy responses.

All chapters feature concrete examples of national policy initiatives, with the aim of contributing to the process of international policy learning. Complexity and uncertainty characterise many aspects of the relationship between developments in STI and the economic and social challenges faced by countries at all income levels. Consequently, the need is ever greater for the exchange of information on policies and the factors found to condition their successes and failures.

References

Bloom, N. et al. (2017), “Are Ideas Getting Harder to Find?”, NBER Working Paper, No. 23782, National Bureau of Economic Research, Cambridge, MA, http://www.nber.org/papers/w23782.

De Backer, K. and D. Flaig (2017), “The future of global value chains: Business as usual or ‘a new normal’?”, OECD Science, Technology and Industry Policy Papers, No. 41, OECD Publishing, Paris, https://doi.org/10.1787/d8da8760-en.

Guillemette, Y. and D. Turner (2018), “The Long View: Scenarios for the World Economy to 2060”, OECD Economic Policy Papers, No. 22, OECD Publishing, Paris, https://doi.org/10.1787/b4f4e03e-en.

IPCC (2018), “Global Warming of 1.5 ºC, an IPCC special report on the impacts of global warming of 1.5ºC above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and efforts to eradicate poverty”, Intergovernmental Panel on Climate Change, Geneva, http://www.ipcc.ch/report/sr15.

Jones, B. (2009) “The Burden of Knowledge and the ‘Death of the Renaissance Man’: Is Innovation Getting Harder?”, The Review of Economic Studies, Vol. 76/1, pp. 283-317, Oxford University Press, Oxford, https://doi.org/10.1111/j.1467-937X.2008.00531.x.

Mazzucato, M. (2018), Mission-Oriented Research & Innovation in the European Union – A problem-solving approach to fuel innovation-led growth, Directorate-General for Research and Innovation, European Commission, Brussels, https://ec.europa.eu/info/sites/info/files/mazzucato_report_2018.pdf.

Miller, D.I. and J. Wai (2015), “The bachelor’s to Ph.D. STEM pipeline no longer leaks more women than men: a 30-year analysis”, Frontiers in Psychology, Vol. 6/37, Frontiers Media, Lausanne, http://dx.doi.org/10.3389/fpsyg.2015.00037.

OECD (2018a), “The policy mix for science-industry knowledge transfer: Towards a mapping of policy instruments and their interactions”, Working Party on Innovation and Technology Policy document, OECD, Paris, DSTI/STP/TIP(2017)7/REV2.

OECD (2018b), “OECD time-series estimates of government tax relief for business R&D”, TAX4INNO Project 674888, Deliverable 2.3: Summary report on tax expenditures, Version 29 May 2018, OECD, Paris, http://www.oecd.org/sti/rd-tax-stats-tax-expenditures.pdf.

OECD (2018c), “OECD review of national R&D tax incentives and estimates of R&D tax subsidy rates”, TAX4INNO Project 674888, Deliverable 3.3: Summary report on tax subsidy rates – core countries, Version 18 April 2018, OECD, Paris, http://www.oecd.org/sti/rd-tax-stats-design-subsidy.pdf.

OECD (2018d), “Research and Development Statistics: Government budget appropriations or outlays for RD (Edition 2017)”, OECD Science, Technology and R&D Statistics (database), https://doi.org/10.1787/e724dc33-en (accessed on 26 September 2018).

OECD (2018e), "General Government Accounts, SNA 2008 (or SNA 1993): Main aggregates", OECD National Accounts Statistics (database), https://doi.org/10.1787/data-00020-en (accessed on 08 October 2018).

OECD (2018f), "Total GBARD (Government budget allocations for R&D) at current prices and PPP", in Main Science and Technology Indicators, Vol. 2018/2, OECD Publishing, Paris, https://doi.org/10.1787/msti-v2018-1-table57-en.

OECD (2018g), “Main Science and Technology Indicators”, OECD Science, Technology and R&D Statistics (database), https://doi.org/10.1787/data-00182-en (accessed on 2 October 2018).

OECD (2017a), Investing in Climate, Investing in Growth, OECD Publishing, Paris, https://doi.org/10.1787/9789264273528-en.

OECD (2017b), The Next Production Revolution: Implications for Governments and Business, OECD Publishing, Paris, https://doi.org/10.1787/9789264271036-en.

OECD (2017c), “Going digital”, in OECD Digital Economy Outlook 2017, OECD Publishing, Paris, https://doi.org/10.1787/9789264276284-4-en.

OECD (2017d), Making Innovation Benefit All: Policies for Inclusive Growth, OECD, Paris, https://www.innovationpolicyplatform.org/system/files/Inclusive%20Growth%20publication%20FULL%20for%20web.pdf.

OECD (2017e), OECD Science, Technology and Industry Scoreboard 2017: The digital transformation, OECD Publishing, Paris, https://doi.org/10.1787/9789264268821-en.

OECD (2016), OECD Science, Technology and Innovation Outlook 2016, OECD Publishing, Paris, https://doi.org/10.1787/sti_in_outlook-2016-en.

Smith-Doerr, L., S.N. Alegria and T. Sacco (2017), “How Diversity Matters in the US Science and Engineering Workforce: A Critical Review Considering Integration in Teams, Fields, and Organizational Contexts”, Engaging Science, Technology and Society, Vol. 3 (2017), p. 15, http://dx.doi.org/10.17351/ests2017.142.

Wilkinson, M. et al. (2016), “The FAIR Guiding Principles for scientific data management and stewardship”, Scientific Data 3, Article No. 160018 (2016), Nature Publishing Group, London, http://dx.doi.org/10.1038/sdata.2016.18 .

Note

← 1.  Findability, accessibility, interoperability and re-use.