The EC attempts to polish the portfolio approach by providing a model addressing the limits and balancing performance across different criteria and objectives.
Decisions play a big role in research funding, whether a project is funded or not goes through a decision-making process based on different criteria and circumstances. Up until now, these decisions have primarily been made through a merit-based approach, meaning that it is done through peer review, and the funds go to the best ranked projects. Yet, some argue that this method lacks flexibility, do not foster unorthodox ideas, and is subjective; thus, not always leading to the best outcomes and undermining quality (see article by Eric Canton).
As an alternative, the portfolio approach is an ongoing experiment put on the table. This approach consists of maximising “an objective function under a set of constraints, using linear or non-linear programming and a heuristic procedure to find maximum values”. It can allow the allocation of funds based on the attributes of sets of proposals rather than funding individual projects independently. This could allow to take into account the interdependencies between proposals and to achieve multiple objectives at the same time. The method is increasingly explored by different stakeholders and the European Innovation Council (EIC) has already piloted the approach. Ultimately, the portfolio would allow to have more synergy and complementarity between projects; therefore, further enhancing the overall impact of projects compared to the sum of each individual impact.
The EC has published a report on the portfolio approach, built on Canton’s article. The recent report complements the first paper by addressing the question of “how such portfolios can be constructed in situations where the number of feasible combinations is extremely large”. Furthermore, it addresses several limits such as those linked to the rule-of-thumb approach, which is largely implemented by funders. While the rule-of-thumb is an easy and simple way to create portfolios, it lacks in taking into consideration the joint performance of the portfolio across multiple criteria; thus, leading to portfolios that perform well in average scores, but unbalanced in relation to policy objectives. Moreover, the robustness of the decisions is weak as it can deliver outcomes that depend on arbitrary thresholds or small differences in rankings. Furthermore, curating proposals is a difficult task because the number of possibilities surges significantly with the number of candidates. In a time where the number of proposals is increasing, this is an unsustainable solution.
In this context, the report provides a new framework for the portfolio approach by tackling the limitations, optimising the performance of the portfolio throughout multiple criteria and by distinguishing high-performing portfolios. The model seeks to directly evaluate portfolios and identify the best possible combinations. It combines a compromise-based objective function with a computationally efficient heuristic search procedure, allowing to create high-quality portfolios without needing to enumerate the full set of feasible combinations. The proposed model is designed in a way that allows flexibility, transparency and adaptability to different criteria and context.
In the competitiveness agenda of the EU, the portfolio approach could help select the projects that can have the maximum impact, therefore, making the most optimal decision and allocation of funding. The framework can enable decision-makers to create well-balanced portfolios that encompass different policy goals under budgetary and organisational constraints. However, the portfolio approach is a highly controversial and contested method, receiving significant backlash from the scientific and academic community. The preference towards merit-based approach comes from long-standing conventions and code of conducts of the R&I sector. Furthermore, although the new framework addresses the limitations of the traditional portfolio approach, it has its own limitations, as different biases and inequalities can also arise. For instance, projects with high scores end up not being funded as it does not fit the overall portfolio and objectives. Although the portfolio topics are not set beforehand, it can reflect the current political agenda and dynamics, rather than the ideas. Not all solutions can satisfy the majority, however, a balancing of values and priorities have to be concretely set to not diminish the core of research integrity and science.