Recent years have seen major efforts to coordinate science and technology policies across Europe. Nevertheless, 87% of public spending on research and development (R&D) in 2014 still remained at the national level.1 This fragmented landscape of domestic R&D support schemes complicates internationally oriented science and leads to a duplication of research efforts within the EU. Increasing the number of joint programming initiatives is therefore a key policy priority for the European Commission (OECD 2012).
The pooling of national contributions to a European programme with a common budget (a so-called ‘real common pot’) poses problems from a governance standpoint. In particular, success rates might be geographically unbalanced if grants paid out to certain countries do not correspond to the country’s contributing share. This debate about net transfers between participants is known as ‘juste retour’ in policy circles. In addition, policymakers might be enticed to lower national funding and to free-ride on partner country’s contributions. Eventually, creating and distributing a single budget at the supra-national level might require changes in the legal framework conditions for participating countries (OECD 2012).
The virtual common pot
To avoid these problems, 80% of European joint programming initiatives are organised as ‘virtual common pots’ (VCPs) (Moretti and Villanova 2012). A VCP emulates a real common pot (RCP) insofar as the quality of projects applying for grants is evaluated centrally by a single public authority, which is responsible for all participating countries. Each country is committed to respect the evaluation ranking of the central agency, but only pays funds to their own national applicants. Consequently, concerns about free-riding and juste retour are mitigated and there is less need to harmonise legal frameworks. However, a VCP makes the allocation process of R&D grants more complicated. Project proposals are usually submitted by international consortia as joint programming initiatives aimed at promoting cross-national R&D cooperation. Thus, budget constraints need to be slack in all countries involved.
To illustrate the working of a VCP, suppose there are four countries participating in a joint programme – A, B, C, and D. Each country contributes a budget to fund exactly two grants. Consequently, there is a total budget of exactly eight grants. Further, suppose that there are project proposals by six international consortia and each project partner requires one grant to conduct the project. All project proposals get evaluated according to a central quality ranking, which may look as follows:
Table 1. The working of a virtual common pot
A consortium formed by two partners from country B and one from country A submitted the best evaluated proposal. Another consortium from countries B and C presented the second-best project, and so forth. In a real common pot, the three highest-ranked applications would be granted, after which the common budget of eight would be exhausted. In a virtual common pot, by contrast, the individual national budget constraints need to be respected. The first proposal receives funding as before. Then, however, country B used up its budget of two and the second-ranked project forgoes funding. For the third proposal, by applicants from different countries, there are again sufficient resources available. Also the fifth-ranked proposal qualifies for funding in a VCP in this example.
Table 1 illustrates three things. First, in a real common pot, no partner from country D receives a grant. Instead, D’s entire budget is paid out to fund partners from other countries. The virtual common pot is supposed to prevent exactly this situation. Second, the VCP leaves gaps of non-funded projects that otherwise would be granted in an RCP. Third, and most importantly, the average quality rank of funded projects is lower in a VCP. Thus, if the impact of grants increases with project quality, less efficient projects are funded. This leads to a trade-off between the geographical balance a VCP is designed to induce and an efficient allocation of the programme’s budget.
The Eurostars Joint Programme
In a recent paper, Hünermund and Czarnitzki (2016) estimate the effect of Eurostars, a joint programme of 33 countries (including five non-EU countries) targeted at R&D-performing small and medium-sized enterprises, on firm growth and job creation. The programme had a total budget of €472 million between 2008 and 2013, of which 25% was co-funded by the European Commission (Makarow et al. 2014).
A practical advantage of the fact that not all high-ranked projects get granted in a VCP is that it facilitates the identification of causal effects. Whereas the best projects always get funded in an RCP, a VCP allows researchers to compare funded firms with their direct neighbours in the evaluation ranking. To put it in econometric terms, the exogenous variation in budget availability, generated by the VCP, serves as an instrument for firms’ subsidy receipt. Hünermund and Czarnitzki use this and show that the impact of R&D grants on job creation is indeed an increasing function of project quality (see Figure 1). Since a VCP funds projects with a lower average quality the estimated relationship translates into one grant-induced job that is 27% more costly compared to an RCP.
Figure 1. Job growth induced by R&D grants in the Eurostars programme depending on project quality
Although a VCP has advantages from a governance perspective, empirical evidence shows that it reduces the efficiency of joint R&D policies. Exact numbers will differ for other programmes, depending on the link between project quality and policy impact. However, there will always be an efficiency loss as long as the relationship is not completely flat. A better trade-off between geographical balance and efficiency might therefore be a combination of an RCP and a VCP. In such a mixed mode a share of the total budget is used to fund the best-ranked projects, irrespective of their geographical origin. The remaining share can still be allocated as a VCP to achieve an evenly distributed granting rate. Hünermund and Czarnitzki simulate a mixed mode for the Eurostars programme and find that if the European Commission’s contribution of 25% had been allocated as a real common pot, a large fraction of the additional costs due to the VCP would have been avoided.
European Commission (2008), “Towards joint programming in research: Working together to tackle common challenges more effectively”, Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions, 11935/08, COM (2008) 468
Hünermund, P, and D Czarnitzki (2016), “Estimating the Local Average Treatment Effect of R&D Subsidies in a Pan-European Program”, ZEW Discussion Paper No. 16-039
Makarow, M, G Licht, I Caetano, D Czarnitzki, and S Elçi (2014), “Final Evaluation of the Eurostars Joint Programme”, Ref. Ares (2014) 3906990
Moretti, P F, and L M Villanova (2012), “Coordinating European national research programmes: the process towards Joint Programming Initiatives”, Consiglio Nazionale delle Ricerche, DTA 10-2012
OECD (2012), “Meeting Global Challenges through Better Governance: International Co-operation in Science, Technology and Innovation”, OECD Publishing