Page 12 - IDEA Studie 07 2023 TACR

ARE SUBSIDIES TO BUSINESS R&D EFFECTIVE? REGRESSION
DISCONTINUITY EVIDENCE FROM THE TA CR ALFA PROGRAMME IDEA 2023
The proposals were evaluated by an expert panel with the help of external reviewers. Each project was assessed by two (calls 1 and 2) or three (calls 3 and 4) external reviewers and one rapporteur from the panel. In the first step, several binary criteria, such as whether the project was within the scope of the programme, were used to eliminate ineligible proposals. In the second step, the evaluators awarded points to each project (0 to 100 in total) based on seven-scale criteria for: i) the research team; ii) expected impacts; iii) market opportunities; iv) collaboration with a research organization; v) topicality and motivation effect; vi) economic efficiency; and vii) the consortium. The projects were then ranked according to the average number of points (score) across the three or four evaluators.
The final decision on whether a proposal was recommended for funding was made by the Board of the Programme and ultimately the Board of TA CR, which had the power to adjust the points, and hence the ranking, for well-founded reasons. However, in practice, the Board of TA CR exercised this power only rarely, for instance, when inconsistencies in a project budget were exposed ex-post. Even in such cases, it almost never happened that a change in the ranking would affect which proposals were actually funded or not. Whether a proposal that met the binary criteria was awarded a grant depended on the amount of funding in a given call.5 The preliminary budgets were known ex-ante, except for call 3, where the funding was unexpectedly increased during the evaluation period. In the other three calls, the bodies involved in the selection procedures had a rough idea of where the cutoff was likely to appear, which raises concerns about possible manipulation of rankings around the cutoff. Because of these concerns, we test the validity of our research design later in the paper using a battery of tests standard in the RD literature.
Table 1 provides an overview of the number of projects in each annual call. In total, 424 proposed projects were subsidised and 1,451 were not. This means that slightly fewer than one in four proposals was funded. The number of proposals increased between calls 1 and 2 and even more between calls 2 and 3, while the number of subsidised projects remained roughly the same; hence, the competition significantly intensified and the success rate dropped in the second half of the programme. The share of proposals
5 Note that various adjustments were made in the evaluation procedures over the course of the programme implementation, especially between calls 1 and 2 and calls 3 and 4. These adjustments, however, did not affect the comparability of the evaluation points across calls. Details of the adjustments are available upon request from the authors.
10