Had they been published at a different time, Amartya Sen¡¯s 1981 book Poverty and Famines and his 1989 follow-up (with Jean Dr¨¨ze), Hunger and Public Action, could have been used to underpin a powerful impact case study.

They broadened public understanding of hunger, promoting a definition that encompasses social as well as biological disadvantages. They challenge the idea that famine is caused by food¡¯s unavailability, instead arguing that it is driven by individuals¡¯ inability to afford it. And they highlight the importance of democracy in preventing famine, arguing for the public provision of goods and services to combat it.

These bold ideas themselves constitute impact, which can be seen in their shaping of the work of national governments, NGOs, the United Nations and many international organisations. They illustrate the kind of bold and adventurous impact that we hope to see submitted later this year to the UK¡¯s latest research excellence framework.

A more recent example ¨C had it been carried out by a university ¨C might be the body of research that led to 2016¡¯s first ¡°¡± to meet the burgeoning demand for mobile services.

ÖйúAƬ

The US auction was the most complicated broadcast spectrum reallocation exercise ever undertaken, pioneering the purchase of certain frequencies from TV broadcasters to resell to wireless providers. To overcome the technical challenges, the Federal Communications Commission and the Smith Institute developed innovative techniques in operations research. In doing so, they facilitated a process that raised nearly $20?billion (?15?billion), saved an estimated $200?million in relocation costs for TV stations and improved services for all mobile services users.

These examples are deliberately drawn from beyond the temporal and geographical remit of the REF to emphasise the need to get beyond rather limited interpretations of what impact means.

ÖйúAƬ

The concept was introduced for the 2014 REF. Yet despite the broad definition adopted, the submitted case studies often interpreted impact rather cautiously. The guidelines for REF 2021 make clearer that impact can be both broad and deep, underpinned by a wide variety of research. The relationship between research and impact is not necessarily linear or direct; more often, significant impact comes from a body of work, effecting a change in understanding.

Hence, institutions should think not only about the types of impact their research has had, but also about the different bodies of research on which it has been founded. The increased weighting for impact in 2021 ¨C rising from 20 per cent to 25?per cent ¨C will appropriately incentivise such efforts. The best impact submissions will likely showcase the full breadth and depth of impacts possible from the equally broad and deep range of research being undertaken in institutions.

In addition, the removal of the standard requirement for four outputs from every submitted researcher, in favour of requiring a fixed number for the overall unit, should better support institutions in investing in the right range of research activity for them. The new requirement to return all staff who have a significant responsibility for research is about promoting inclusion, addressing the negative consequences for staff associated with exclusion in previous exercises, and encouraging the presentation of a rounder picture of research activity.

Far from risking homogenising research activity, as some appear to argue, these changes will give universities greater freedom to pursue research of different kinds ¨C including longer-term and higher-risk research that can bring high rewards. And the ability to submit the outputs of staff who have since left the institution will increase incentives to invest in sustainable research environments.

ÖйúAƬ

These changes also underline the move away from the previous focus on individuals, on to the research produced by groups of researchers in the submitting unit. Requiring a fixed number of outputs for the overall unit will better reward collaboration, give institutions flexibility in building submissions and better recognise the different lives and personal situations of researchers.

Neither staff names nor the number of outputs attributed to them will be published by the funding bodies, but institutions must follow this lead and resist the temptation to use the number of outputs submitted for a given researcher as a marker of status.

Finally, analysis from REF 2014 highlights the sheer volume of impact case studies that were underpinned by interdisciplinary research, even though they were submitted through a particular discipline. This underlines the case to ensure that such research is properly recognised, appreciated and rewarded and is why the four UK ÖйúAƬ funding bodies have appointed a specialist advisory panel to guide and oversee assessment processes. Moreover, interdisciplinary advisers have been appointed to both main panels and subpanels to ensure that such research is reviewed by those with the right expertise.

We hope that this assurance of its fair and equal assessment will encourage institutions to submit interdisciplinary work to REF 2021. As with impact, it will be valuable to have the bold, the adventurous and the diverse well represented within the national research profile.

ÖйúAƬ

We trust that these changes will mark a new and productive phase in fostering the deep, creative and high-impact research in which the UK is so strong ¨C and can be still stronger.

Lord Stern is IG Patel professor of economics and government at the London School of Economics. He was president of the British Academy between 2013 and 2017 and chaired the 2016 REF review. David Sweeney is executive chair of Research England.

ÖйúAƬ