Point of no returns: researchers are crossing a threshold in the fight for funding

Million-dollar questions

My main research focuses on using computational tools to analyse intelligent systems, but I have found myself increasingly questioning how the ways in which we fund science shape its outcomes1. Which funding schemes encourage researchers to pursue high-risk research? How do different funding schemes affect scientists themselves, and what ethical issues arise from them?

I suspect that because competition had been a constant companion on my path to becoming a professor at the Vienna University of Technology (perhaps even paving the way), I developed a particular interest in analysing its implications. These questions are core topics in metascience, which takes a bird’s-eye view of how research is done and aims to improve its quality, integrity and efficiency. Like many of my colleagues, I work on these topics alongside my main research.

A concept known as the Szilard point helps to contextualize the issues arising from excessive competition for grants. Named after the Hungarian-born physicist Leo Szilard, who wrote a short story satirizing the bureaucratic nature of scientific funding, this metric describes the threshold at which the total cost of competing for a grant equals (or surpasses) the value of the available funding. These costs are incurred by scientists in writing proposals, by their peers in reviewing them and by the administrative systems that run the process. The question is, which costs more: the research being funded, or the application process itself?

GenAI for Africa, a funding call from the European Union’s Horizon Europe programme, probably crosses that threshold. The initiative aims to use generative artificial intelligence to address the societal challenges that many African countries are facing. With a total budget of €5 million (US$5.8 million), the call, which closed in October, invited proposals across four vast domains — agriculture, health care, urban planning and education. Out of 215 submissions, only two projects are expected to be funded, giving a success rate of under 1%.

To approximate the overall costs associated with the application process, I used two scenarios. See ‘GenAI for Africa: estimated grant-application costs’.

Scenario A

In 2023, group of researchers at the University of Lübeck in Germany developed a simulation tool for estimating the costs of grant funding. To arrive at cost estimates for GenAI for Africa applications, I fed this simulation with data from the Interim Evaluation of the Horizon Europe Framework Programme for Research and Innovation (2021–2024). The key inputs are:

Time investment. According to the report, “the median consortium coordinator spends … 36 to 45 person-days per proposal. The effort for contributing consortium partners is typically lower, spending 16 to 25 person-days.”

Consortium size. The average consortium size for Horizon projects is likely to be between 12 and 16 partners, according to statistics published by the European Commission.

Hourly payment rate of applicants. Hourly rates vary considerably depending on country, sector (industry versus academia), seniority and applicable overheads. To reflect this diversity, I drew on several reference sources from across Europe. To account for differences, I assumed the average cost of an hour’s work to be between €20 and €60.

The variation in input values enabled me to produce two cost estimates for GenAI for Africa grant applications: a lower estimate (Scenario A1) and a higher estimate (Scenario A2).

The tool automatically accounts for decision-making and administrative costs.

Scenario B

The evaluation report of the Horizon Europe Framework offers an alternative estimate, indicating that an average proposal costs between €21,000 (Scenario B1) and €32,000 (Scenario B2). The costs for the decision-making process and administration were added on top; I estimated that 85% of total costs fall on applicants, 10% on reviewers, and 5% on administration2. I then summed up the costs across all 215 applications received in response to the GenAI for Africa funding call.

My estimates of the cost of applications range from €5.3 million to about €43.5 million. Even the lower estimate exceeds the €5-million value of the GenAI for Africa grant, making it very likely that this scheme has been a net drain on the research system.

In other words, European taxpayers will have spent more on the funding process than on the funding itself, and the scientific ecosystem has been drained.

(A Nature editor contacted the European Commission for comment on this article, but had not received a response by the time of publication.)

How can funders avoid crossing the Szilard point?

Avoid overly broad funding calls. When multiple disciplines compete for a limited pool of funds, the probability of success collapses. Agencies must create focused funding calls that target specific research areas. Funders might want to seem generous, and supportive of all areas of science, but only those with the deepest pockets should be announcing broad calls.

Reduce opportunity costs. Staged application procedures, in which short initial proposals determine who is invited to submit a full proposal, seem to be more cost-effective than single-stage systems.

Experiment with other models. A German study3 published last month in Nature Communications tested lottery-first approaches, in which a draw decides who may submit a full proposal. The results showed that these systems cut costs by roughly two-thirds compared with conventional competition. Democratic voting4 processes, whereby scientists name peers whom they consider deserving of funding, can also greatly reduce the costs associated with grant applications.

Research-funding systems profoundly shape how science operates, and it is frustrating to encounter schemes, such as GenAI for Africa, that seem to end up being a waste of time and money. Scientists should expand public knowledge and tackle global challenges. They could do this much more effectively with better funding systems.

This is an article from the Nature Careers Community, a place for Nature readers to share their professional experiences and advice. Guest posts are encouraged.

Subscribe to Nature Briefing: Careers, an unmissable free weekly round-up of help and advice for working scientists.