Direct supports to innovation in companies in the form of grants, loans and subsidies constitute a key feature of Swedish and European innovation and enterprise policy. Unlike for example general tax credits, direct supports require that authorities actively select, and thereby deselect, which research and innovation projects to fund. The question immediately arises how such a selection should be conducted to ensure distributed funds have the desired effect and that policy instruments attain their goals.
International research as well as reviews of Swedish innovation support schemes emphasize that implementation and design of selection processes are crucial and that every such process contains both pitfalls and difficult tradeoffs. Well-functioning selection processes are consequently of high relevance for the implementation of the Government’s export, reindustrialization and digitalization strategies which include direct supports.
This literature review is a part of the Swedish Agency for Growth Policy Analysis’ (Tillväxtanalys) government assignment on methods and models for evaluating and analyzing public interventions. The report will form a basis for learning seminars for those authorities engaged in distributing direct supports for businesses and innovation. The literature review, alongside an upcoming empirical study, aims to provide an understanding of different selection processes in the field, their strengths and weaknesses and when they can be expected to work satisfactorily. The focus of this review has been on identifying the common difficulties, pitfalls and tradeoffs that authorities have to deal with and we have found the following:
The logic of direct support schemes is that not every business or project needs every type of support. Efficient policy is based on the ability to select recipients with accuracy. But how is the target group to be separated from the general population of firms or projects? This requires diligent work to delimit the target group and define eligibility already in the drafting of the support scheme, as well as working to in practice determine which firms or projects fall within the set limits. The work is made more difficult by the tendency for the target group to be the least aware of the availability of support and to be most limited in their capacity to apply. Efforts to inform and encourage the target group to participate is therefore of importance.
A major challenge for those authorities distributing direct supports is to determine additionality and utility in a situation defined by asymmetric information – firms know more about their projects and their actual need of support – and misaligned incentives – firms have an interest to represent themselves as in need of support and their project as having great social utility. The authorities’ processes therefore need to guarantee that the right type of credible information is provided to the decision maker.
Another difficult challenge for authorities administrating direct support schemes is to secure the right expertise and to delegate each decision to the right expert. If decision makers do not have the right expertise they will often fall back on informal decision criteria and processes that do not match the stated program logic. In such cases where the authorities do not themselves have the necessary expertise they might need to rely on external experts. This entails a risk for agency problems, including misaligned incentives, putting high demands on guidelines, clear selection criteria and training.
Finally, decisions to hand out direct support are made by a relatively large number of individuals with differing backgrounds, incentives and opinions. In addition, the decision makers need to consider several decisions simultaneously. Putting together a final decision from multiple sources gives rise to its own problems and different processes will lead to different decisions. There is, for example, a tradeoff between conservative but thorough decisions on one hand and pluralism and more radical projects on the other.
The literature which we have found and reviewed here is largely based on single direct support schemes, rather than comparison, and positivist rather than constructive. In other words, it brings out what has gone wrong in some instances but has less to say about how to do selection right. It is consequently valuable to analyze how authorities have handled selection problems and difficulties in their specific contexts. For example, the selection process must be modulated by the size of the support. If a single support is small a very expensive but accurate selection process is unnecessary, whilst being crucial for larger supports. Some support schemes have a regional focus, which demands a degree of decentralization and delegation. Support schemes can also vary by technological focus, which affects the need for external expertise. Furthermore, some schemes face differing application rates which affect, for example, how efforts to inform the target group should be structured. If insights from the literature are to be applied in full they need to be adapted to the specific context of individual support schemes.
While it isn’t possible to design the ideal selection process based on the existing literature, it is possible to draw some general lessons about how authorities can work with selection processes. The first such lesson is that selection effects, meaning that the firms that apply are different from the ones that do not, frequently undermine the possibility to evaluate direct support schemes. It’s often impossible to say whether any measured effects, positive or negative, are due to the support or the selection. Authorities can therefore improve possibilities for evaluation by proactively analyzing both the target group and the broader population of firms to enable comparison of those who apply with those who do not.
It is clear that many steps of a selection process put high demands on decision makers knowledge of problems that could potentially arise. The importance of systematic and specialized training is raised time and again in the literature. This insight is closely connected to the need for institutional memory and documentation. Problems which have arisen in previous iterations of a scheme can be avoided if insights are transmitted and incorporated in both implementation logic and training.
The literature puts great weight on the interplay between all steps of the selection process. A strategy for identifying the target group is closely linked to the information to be demanded from the applicants, which in turn is linked to the expertise needed to process that information in a rational way. Selection and the different steps of the process therefore cannot be an afterthought but must harmonize with the program logic already at the support scheme design stage.
The upcoming empirical study will depart from the common problems described here and collect relevant information about how authorities deal with difficulties and tradeoffs based on their respective contexts. The goals are to (i) describe and contrast three different but representative selection processes at Vinnova, the Swedish Energy Agency and the Swedish Agency for Economic and Regional Growth, (ii) discuss how the authorities handled difficulties and tradeoffs, (iii) connect authorities’ discussions about selection processes to existing theory and (iv) depending on data availability analyze the effects of selection process design on selection outcomes.
Serial number: PM 2018:05
Reference number: 2017/133