Officials in the European Commission are preparing the ninth research and innovation programme and a preliminary proposal, “will be on the table by the end of 2017, beginning of 2018” as part of the Commission's long-term budget planning, said Robert-Jan Smits, Director-General for Research and Innovation, speaking at the Science|Business Horizon 2020 conference this week.
“We are launching a foresight exercise to [establish] the big societal challenges we should focus on. We have commissioned a couple of think tanks to help us and we’re already seeing topics emerge [which are] not in Horizon 2020,” Smits said.
Smits noted this new approach to preparing a research programme was suggested by the eminent Belgian microbiologist Peter Piot, co-discoverer of the Ebola virus.
The initial study is a prelude to the public consultation, which will seek input from academics and industry on the research the EU should fund in the long term. The Commission’s calendar sets 2017-18 for a rethink of its budgetary planning – the so-called Multiannual Financial Framework, of which Horizon 2020 is a part. That planning is expected to contain the orientations and budget for Framework Programme 9, Commission officials said after the conference.
Horizon 2020 continues to be inundated with grant applications, with 90,641 received in a little over two years. This is “a daunting thing” and means the success rate is stuck between 12 and 14 per cent. “I’m afraid it could remain around that level,” Smits said.
While not optimistic of success, Smits has asked for more money in advance of an EU spending review. “Did I put in a claim for more money? Absolutely. Am I going to get it? I think it will be difficult,” he said.
In lieu of a bigger budget, some of the programme rules are being amended. “We’re enlarging two-stage procedures across the board,” said Smits. As a rule of thumb, 80 per cent of proposals, which are not good enough to make the cut, should be rejected in a short-form, stage one evaluation. In stage two, where a longer application is required, at least 35 per cent of proposals should have a chance of success.
Smits reiterated that competitions will not become more rigid, which would have a natural effect of discouraging some applicants. “In [Framework Programme 7], we had very narrowly defined topics, he said, noting that in Horizon 2020, “We went away from a very prescriptive approach.”
There is a reluctance to adopt the European Research Council method, where applicants who score below a certain threshold cannot apply again for two years. That is, “Easy to do for an individual grantee, but you can’t do it with a consortium,” Smits said.
Applicants must to be selective. The inclination to send a proposal to Brussels might occur suddenly, “If it’s Sunday, it’s raining outside and the kids are screaming.” But, Smits advised researchers, “think twice about it”.
Despite Smits’ soothing words, there is discontent amongst scientists about the Horizon 2020 evaluation procedure.
“Many say evaluation results are less predictable than before,” said Fabrizio Gagliardi, chair of the Association of Computing Machinery.
Typically applications are scored by three evaluators working in isolation of one another. If opinion is split on a proposal, the scores of the three evaluators diverge, and it is not clear to applicants how the final decision gets made. Several delegates at the conference asked for a return to the old days, when reviewers held consensus meetings to challenge each other’s scoring. This feature wasn’t included in Horizon 2020 because it slows down time to grant.
Massimo Busuoli, head of the Norwegian University of Science and Technology’s Brussels office, said divergence among reviewer scores was confusing for applicants. Sometimes only one out of the three evaluators will score a proposal down, but will carry the day even if the other two reviewers were positive. “In many cases, the opinion of one expert can change the other two opinions. The process should be more transparent,” Busuoli said.
Caroline Bergaud, managing partner of Bergaud & Partners, and an evaluator for some past SME Instrument and Fast Track competitions, said that to some extent evaluation results are a lottery. “It is related to the fact that we can’t attract enough talent from companies. Evaluating the evaluators would be a good idea,” she said.
Scientists also complained about meagre feedback from evaluators, another time-saving element, which can leave failed applicants in the dark.
“Ensuring quality is a huge challenge for us,” Smits conceded. But he flatly disagreed with Bergaud’s comment that some evaluation processes turned on chance. “Nobody involved would call it a lottery; we always go for quality,” he said.
To become an evaluator, scientists and industry representative are invited to register themselves to an online database of experts. Evaluators are then chosen using a keyword search.
In the past experts were hand-picked, Smits noted. “I’ve also come across people when I was involved in evaluations who call themselves experts but are not really. I once had an evaluator who said he was fluent in English. The person did not speak a word of English and then you’re stuck with them for a week.”
Another problem is the preponderance of scientists among evaluators. It is difficult for the Commission to recruit evaluators from the private sector. “Academics [can be] free for two or three weeks,” Smits said. But experts from private sector “usually cancel at the last minute” for reasons related to workload or because the incentive to be an EU grant evaluator is not that strong.
The midpoint review of Horizon 2020, scheduled for later this year, is a time to consider shifting some furniture – although Smits advised against changing too much.
In terms of the timetable for the new European Innovation Council (EIC), which EU Research Commissioner Carlos Moedas is keen to progress quickly, Smits said, “I think we have to first see the results of the EIC consultation,” expected in May.
“We have enough flexibility inside the last three years to shift things; to change things in work programme 2018, 2019 and 2020, if we consider it necessary to shape the EIC,” said Smits.
However, he added, “I would be very hesitant to break open the legislative procedure for Horizon 2020.” Revising Horizon 2020 would require approval from the European Parliament. “This can take up to two years,” Smits noted. “There’s so many different interests in the Parliament. For Horizon 2020 for example, we received over 2,000 amendments. It was very delicate.”