The coronavirus (COVID-19) crisis presents exceptional challenges across all aspects of public policy. Within the context of science, technology and innovation (STI), governments are increasingly seeking to draw on society’s full potential for innovation, using collective intelligence. By efficiently harnessing the knowledge and expertise of groups, and focusing this on specific research, innovation and policy problems, such approaches could help governments address the current pandemic, as well as future crises. This brief provides an overview of several approaches, including innovation prizes, prediction markets; and, support for open source and other online engineering projects. All of these approaches could complement existing STI policies.

Although prizes have been used since the 18th century to generate innovative ideas, web platforms are enabling governments, corporations, and charities to operate new types of innovation prizes. Compared to grants, innovation prizes are particularly well-suited to tackling many problems associated with COVID-19, notably in areas where success is relatively easy to define, the range of potential contributors is wide, and there is little prior knowledge of who is most likely to make breakthroughs.

Prizes could be used across many topics relevant to COVID-19, from devising tests, remedies, and cures, to innovations in social distancing. One such initiative, the COVID-19 Open Research Dataset Challenge (CORD-19), was launched in the United States in March. Sponsored by the Allen Institute for AI and eight partner institutions, CORD-19 offers cash prizes for the best responses to critical research questions examinable using a machine-readable data set of 44 000 scholarly articles. In another example, Montreal’s General Hospital Foundation has sponsored the Code Life Ventilator Challenge. Open to groups around the world, the CAD 200 000 prize aims to elicit multiple designs for inexpensive, safe and effective machines, some of which might also be suited for manufacture and use in other parts of the world.

Prizes should be designed in a way that incentivises individuals or research teams to share information widely and rapidly. For example, the Defense Advanced Research Projects Agency (DARPA) Network Challenge used financial incentives to both encourage participation in the challenge, and to reward participants for recruiting others to the same task. Similar principles might be employed to facilitate information flow in new innovation prizes. In April 2020, the Joint European Disruptive Initiative (JEDI) announced its first-ever DARPA-style challenge to help fast track a therapeutic treatment for COVID-19. JEDI’s Billion Molecules against COVID-19 Grand Challenge will offer prizes of up to EUR 2 million for winning teams.

Recent empirical work also finds that winner-takes-all compensation schemes (for individuals or teams) generate significantly more novel innovation than those that offer the same total compensation but are shared among multiple winners. However, prizes will only work in some areas; they tend to be less effective and efficient than other types of incentives if considerable effort is required to formulate the challenge and validate, test and implement proposed solutions.

Another mechanism that could help inform policy responses to COVID-19 is a prediction market (also known by such terms as “information market” or “decision market”). Prediction markets have outperformed experts in predicting outcomes in fields as diverse as sporting tournaments and political elections. In general terms, a prediction market provides a mechanism for individuals to bet on whether a specific outcome will occur in the future. Prediction markets can now be established quickly and inexpensively through specialised digital platforms such as Gnosis and Augur.

Like markets more generally, prediction markets gather decentralised private information from market participants, who may be many and unknown to each other. They also incentivise participants to discover new information about the prediction, because better information could be profitable. For instance, a prediction market might be established for a testable scientific hypothesis. Participants would place bets until the hypothesis is proven or disproven, with every participant having a financial incentive to gather (new) information about the validity of the hypothesis. Over time, if bets converge around a particular result – a true or false hypothesis, in this case – the potential financial gain from subsequent bets on the same result falls (as when the odds shorten for bets on an increasingly likely sports result). The market makers (in this case public agencies) can see the information being provided through the changing prices (or odds) after each new bet. In the science domain, recent experiments show that prediction markets could help in such tasks as:

  • Predicting the results of otherwise expensive research evaluations (e.g. of higher education institutions).

  • Quickly and inexpensively identifying research findings that are unlikely to replicate.

  • Helping optimally allocate limited resources for replications.

  • Helping institutions assess whether strategic actions to improve research quality are achieving their goals.

  • Helping understand specific scientific processes. For instance, a research project could be examined alongside a history of the project’s market prices, to show when hypotheses had strengthened or weakened.

In policymaking, prediction markets remain underused. In the context of the current crisis, prediction markets could help policy makers in many ways, from assessing the utility and timing of social distancing measures, to forecasting the disease’s progression (so long as there is a measurable outcome to bet on). Prediction markets could complement a portfolio of other information sources. Human intelligence (of individuals or crowds) and machine intelligence might also be combined. For instance, research is underway on using artificial intelligence (AI) systems in prediction markets, with the AI placing bets alongside humans based on its own forecasts.

From the start of the COVID-19 crisis, a number of governments have offered grants or issued calls to industry to increase production of urgently-needed medical equipment, some of which involve significant technical challenges. Some ventilators, for example, have as many as one million lines of software code and over 1 700 parts.

Several open source collaborations have sprung up to meet this challenge. HelpfulEngineering is an open-source collaboration with over 3 000 members working on ventilators, oxygen concentrators and other medical equipment, and the maker movement (a technology-oriented community of build-it-yourself enthusiasts) is active across many COVID-19 initiatives. The MIT Emergency Ventilator (E-Vent) Project aims to identify the minimum requirements for a low-cost ventilator based on the collective know-how of clinicians, and plans to design, test and report against these requirements. One initiative, in the United Kingdom, has come up with the “exovent”, which dispenses with the ventilator entirely, instead employing a mechanical capsule around the thorax, avoiding medical complications that intubation can cause as well as the need to sedate the patient.

Governments could benefit from connecting to such open-source initiatives to assess what support the projects may need from which public authorities, and how projects might best align with government priorities. Governments could also diffuse information about the projects and offer strategic direction, where relevant, help drive consensus on optimal designs, and highlight information on acceptable clinical specifications for machines and components. As these types of initiative may not have sufficient time for crowdfunding or more formal early-stage equity investment amid a crisis like COVID-19, governments could offer emergency funding to build prototypes.

As part of short-term responses to the COVID-19 crisis, governments could:

  • Establish a spectrum of innovation prizes across key topics relevant to the COVID-19 response, creating new prizes as challenges emerge, such as optimal planning of return-to-work scenarios. Prizes might also be considered for topics of specific relevance to low-income countries.

  • Establish a broad and rolling set of prediction markets across topics relevant to the COVID-19 crisis.

  • Design innovation prizes to emphasise information sharing. If multiple government agencies are deploying prizes that target similar problems, the merits of pooling resources could be assessed.

  • Provide additional funding to help elicit and develop certain types of proposals (for instance those requiring costly prototypes), because liquidity constraints can limit some firms’ participation in prizes.

  • Connect to open-source and other engineering projects to assess how projects might align with government priorities (for example, developing low-cost ventilators) and determine what support the project teams may need from public authorities.

  • Diffuse information about open source projects and offer strategic direction where relevant. Governments might also help to highlight information on acceptable clinical specifications for machines and components and help drive consensus on optimal designs.

  • Offer emergency funding to build prototypes for open source medical equipment, as there may be insufficient time for crowdfunding or more formal early-stage equity investment.

Further reading

OECD (2020), The Digitalisation of Science, Technology and Innovation: Key Developments and Policies, OECD Publishing, Paris,


This paper is published under the responsibility of the Secretary-General of the OECD. The opinions expressed and the arguments employed herein do not necessarily reflect the official views of OECD member countries.

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area.

© OECD 2020

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at