Public Support to Business Research and Development in Belgium - Fourth evaluation [ REP_12721 - 22/11/2022]

  PDF & Download

• Publication (en - 8.48 Mb)

  Authors

Michel Dumont (A)
 
A : Author, C : Contributor

  Publication type

Reports

The report presents the work carried out by the Federal Planning Bureau in its fields of expertise, at the request of the authorities, partners. 

Executive summary

Belgium was one of only seven EU countries to reach its Europe 2020 Research and Development (R&D) investment target in 2020. With R&D expenditures equal to 3.48% of Gross Domestic Product (GDP), Belgium amply exceeded its 3% R&D intensity target and took second place, narrowly behind Sweden. The high R&D intensity in 2020 is partially explained by a Covid-19-induced drop in GDP (the denominator of R&D intensity), but Belgium is nevertheless one of the EU countries with the highest increase in R&D intensity since 2010.

In fulfilment of its commitment to the 3% target, set at the 2002 Barcelona European Council meeting, the Belgian federal government introduced several tax incentives in support of business R&D, starting from 2005. This indirect support through tax benefits supplemented the existing substantial direct support (subsidies), provided by the three Belgian regions. The R&D tax incentives were provided through personal income taxation, such as a partial exemption from payment of the withholding tax on the wages of R&D personnel, or through corporate income taxation, such as a tax credit for R&D investment and a patent income deduction. In 2018 a partial exemption for R&D personnel with a bachelor’s degree was added to the four existing partial exemption schemes. The patent income deduction, introduced in 2008, was replaced in 2016 (phased out in 2021) by an innovation income deduction, developed according to the Base Erosion and Profit Shifting (BEPS) guidelines of the OECD, which aim to tackle harmful tax avoidance by multinational enterprises. The tax incentives for business R&D have gradually gained popularity and the estimated budgetary cost consequently increased substantially, reaching a total of 2,782 million euro in 2019 (0.59% of GDP), especially due to a strong increase in the cost of the corporate income taxation benefits.

This report presents the results of the fourth evaluation of public support to business R&D in Belgium, which aims at providing an indication of the extent to which direct support (regional subsidies) and indirect support (tax incentives) have contributed to the strong rise in R&D intensity, by supporting R&D activities of companies that would not have been performed without public support. The assessment of the effectiveness and efficiency of public support is complicated by data limitations and the difficulty of estimation procedures to provide an indication of the causal impact of support on the R&D expenditures of companies. In keeping with the previous evaluations, the estimation strategy adopted in this evaluation is to consider a baseline estimation and to compare the results with the results from alternative estimation procedures, to establish whether robust conclusions can be obtained.

The evaluation provides robust indications that direct support (regional subsidies), and the partial exemption from payment of the withholding tax on the wages of R&D personnel, encourage companies to invest in R&D activities, in addition to the public support that they receive. This result confirms the conclusions of the previous evaluations. This finding also holds for the partial exemption for R&D employees with a bachelor’s degree, which was not included in previous evaluations as it was only introduced in 2018. Equally in line with previous evaluations, there are few indications of additionality for the tax credit for R&D investment and the patent income deduction. The tax deduction for R&D investment, which given partial information was not fully assessed in the previous evaluations, is found to result in additional R&D expenditures by companies. The most worrying finding of this fourth evaluation concerns the innovation income deduction. This corporate income taxation incentive was introduced in 2016 to replace the patent income deduction. The estimations presented in this report provide robust indications of crowding out for this tax scheme, that is, it appears that the innovation income deduction is financing R&D expenditures that companies would finance themselves in the absence of the tax support. The fact that the corporate income taxation incentives, except for the tax deduction for R&D investment, seem ineffective or even result in crowding out, points at an opportunity to increase the efficiency of R&D tax benefits, especially considering that they claim the lion’s share of the budgetary cost of public support to business R&D in Belgium.

This report also provides estimates of the impact of the innovation bonus, which consists in a compensation that is exempted from social security contributions, for workers that generate innovative ideas within a company, and EU funding of research by Belgian companies. Estimates suggest additionality for the innovation bonus but crowding out for EU funding.

Considering distinct groups of firms and industries along several dimensions, reveals substantial heterogeneity in the impact of public support, even with opposite signs for different groups of firms and industries, that may cancel each other out in estimations that consider all R&D active firms as a homogenous population. For example, it appears that the crowding out of some corporate income taxation incentives mainly applies to large and older firms, firms that belong to a multinational group, and to highly concentrated industries. These results should draw attention to the potentially negative impact of public support on market dynamism, as it may reinforce market concentration and winners-take-most effects, and to low efficiency of public support in highly concentrated industries. The results in this report reveal substantial heterogeneity in the impact of public support across industries, but even more so across firms (within industries). This suggests that by targeting specific industries or groups of firms, the effectiveness and efficiency of public support may be increased. Such an approach however requires a well-defined and evidence-based framework, which does not appear in prospect today. Moreover, the conditionality of public support may be at odds with EU state aid rules, which generally prohibit public support to specific companies or industries, although this is right at the core of the current discussion on industrial policy and mission-oriented programs.

In line with the previous evaluations, there are clear indications that when companies combine several public support schemes, the effectiveness of individual instruments in stimulating R&D decreases substantially. However, the combination of different support instruments does not appear to be the problem as such, but rather the combination of large amounts of support without any cap on the total amount of public support that companies receive. The effectiveness of public support decreases with increasing total public support, both in terms of the rate of support and the total amount of support. The crowding out of corporate income taxation incentives is revealed at the highest levels of the total amount of public support. This suggests that the introduction of a cap on public support can contribute to an increase in the efficiency and can be instrumental in containing the considerable rise in the budgetary cost of public support, which is predominantly due to those tax incentives that appear least effective. An analysis of 20 OECD countries, including Belgium, indicates that R&D tax incentive schemes that cap the amount of supported R&D expenditure, or reduce the support rate once a certain threshold has been reached, are likely to show greater additionality.          

Arguments for public support to business R&D rely mainly on the assumed existence of a positive impact from the R&D activities of companies on the rest of the economy. These spillovers create a gap between the private return to R&D and the social return to R&D. As companies are only interested in the private return of their R&D activities, they may not invest sufficiently from a societal perspective, hence the potential role of subsidies and tax incentives to support business R&D. This report examines, in more detail than the previous evaluations, the role of public support in the potential results of R&D activities. The choice of output indicator, and of variables through which spillovers may be detected, is however not trivial and hampered by the lack of a clear indication as to which output indicators law makers had in mind when introducing public support. Considering indicators like productivity, turnover, value added and profit, self-financed R&D seems to generate a positive private return. The private return to R&D financed with regional subsidies appears to be even higher than self-financed R&D. The private return to R&D financed through partial exemption from payment of the withholding tax is found to be negative in some cases, which could indicate that R&D activities financed through these schemes support marginal activities. Estimates also suggest that R&D financed through corporate income taxation incentives generates a positive return for the beneficiaries of the support, though generally lower than the return to self-financed R&D and R&D financed with regional subsidies. However, as these tax incentives can only be used by profitable firms and firms with successful past R&D activities, resulting in current patent or innovation income, estimates of a positive “impact” on output may simply reflect that good performance is a necessary condition to benefit from these tax incentives. There are indications of positive spillovers, for example from R&D of firms that combine support schemes, but also of negative spillovers which may hint at business-stealing effects and imitation by laggards. Young firms, domestic firms that do not belong to a multinational group, and firms with only occasional R&D activities, do not appear to benefit from R&D by other firms. This casts doubt on the extent to which knowledge spills over to Belgian companies that do not belong to a multinational group. Because the necessary data are currently not available, this evaluation does not consider foreign R&D, which is known to be an important source of spillovers for companies in small open economies like Belgium. The absence of foreign spillover variables may bias the estimates of the private return to R&D, and the estimates of domestic spillovers. The inclusion of foreign spillovers in future evaluations is certainly worth considering. Although the potential results of business R&D are the ultimate motivation for public support, the mentioned limitations in estimating the impact of R&D financed through public support on the output of the beneficiaries of support as well as on the rest of the economy, warrant substantial caution in the interpretation of the estimates on output presented in this report.    

Recently, several authors and international organizations have started to warn for the risk that public support may result in the over-subsidization of applied research and the under-funding of (public) basic research despite the well-known importance of basic research and the complementarity between public and private R&D. Belgium has become one of the most generous OECD countries in terms of tax support to business R&D. As tax incentives tend to encourage applied research and experimental development, more than basic research, a reflection may be appropriate on whether the mix of public support in Belgium does not overly encourage applied research and experimental development, at the expense of investment in (public) basic research and complementarities between companies and other actors of the innovation system.