The Engineering and Physical Sciences Research Council is to decrease the share of its portfolio allocated to medical imaging, superconductivity and seven other areas, according to its latest prioritisation exercise.

The results of Balancing Capability, one of the council’s three funding strategies for 2015-20, were published on 15 February. Of the 111 areas in engineering and physical sciences, EPSRC expects 12 will attract more funding as a proportion of total spending, nine will see a smaller share, and 90 will retain the same proportion.

The research and training portfolio for medical imaging, the only area in which EPSRC has provided numerical estimates, has been projected to decrease from £92 million to £73m over the next five years.

In an interview with Research Fortnight in October 2016, the council’s deputy chief executive Tom Rodden said that the “envelope of growth or shrinkage” for each area would be no more than 2-3 per cent.

Areas that are expected to see funding boosted include energy storage; materials for energy applications; robotics; natural language processing; statistics and applied probability; electrical motors; and pervasive and ubiquitous computing—aimed at harnessing the opportunities of the Internet of Things, connected cities and digital health.

Both Rodden and EPSRC chief executive Philip Nelson stressed that in developing Balancing Capability, the council has learned the lessons from its 2012 Shaping Capability exercise that came under fire from researchers. Seventy seven scientists—including Nobel laureate Harry Kroto—signed an open letter saying that “unqualified EPSRC staff” were making decisions on funding without consulting researchers and called for the council to be overhauled.

Nelson said that this time, the EPSRC has spent 18 months consulting. The process included a public call for evidence, which closed in June 2016 and received more than 1,000 submissions. About 70 universities, businesses and professional bodies participated.

The evidence collected, along with input from the learned academies, fed into a list of ‘rationales’. These are five-year strategies for each of the 111 areas, developed by the EPSRC’s strategic advisory teams and a special Strategic Advisory Network created for this exercise. Final decisions were made by the EPSRC’s governing body at its December 2016 meeting.

Nelson insisted that excellence remains the top criterion in assessing all applications, and that Balancing Capability will be considered only when two excellent proposals are ranked equally.

“Researchers in a ‘reduce’ area should not get overly worried because ‘reduce’ does not mean that we will stop funding that research. It would be sensible for them to look at the broad portfolio and see if they can contribute to any adjacent area, but it is important that they don’t overreact and stop doing good work which still has potential to be funded,” Nelson said.

Mark Blamire—head of the University of Cambridge’s materials science department and winner of a £2.7m grant from the EPSRC in 2016 to study superconductivity—expressed concerns about the impact of the exercise. Receiving the ‘reduce’ label, he said, may lead to a sharper decrease in the number of proposals on superconductivity being submitted, causing a faster contraction than the EPSRC had foreseen.

“If responsive mode is supposed to mean what it says, then it is unnecessary to undertake this exercise. The EPSRC might have put in a lot of work, but it is not obvious that it is of any particular benefit to anyone.”

The Royal Academy of Engineering, the Institute of Physics and the Academy of Medical Sciences declined to comment on the EPSRC’s decisions by the time Research Fortnight went to press. Nelson denied that the exercise was risky. “We are not slavishly wedded to this direction of travel and we will absolutely change course if there are good reasons to do so,” he said.

Nelson said that he hoped the EPSRC would continue to carry out the exercise “in one way or another” once the council becomes part of UK Research and Innovation.

This article also appeared in Research Fortnight