The effects of the R&D deduction on turnover, value added and labor productivity
In 2014, Sweden introduced an R&D tax incentive aimed at strengthening firms’ incentives to employ researchers and thereby promote innovation and growth. Unlike traditional tax deductions linked to profits or income, the Swedish R&D tax relief takes the form of a reduction in employers’ social security contributions for R&D staff and can be utilised immediately.
This design particularly benefits labour-intensive R&D and start-ups that are not yet profitable. Moreover, the Swedish scheme includes a cap that limits the total subsidy per corporate group, which is advantageous for small and medium-sized enterprises (SMEs).
A previous study by Growth Analysis (WP 2022:1), later extended in an academic paper by Stavlöt and Svensson (2025), found that the reform had significant effects on the employment of R&D and STEM staff and that it was widely used by smaller firms. Over time, increased R&D activity is expected to contribute to higher productivity, strengthened innovation capacity and long-term economic growth. This report examines to what extent the increase in R&D employment has translated into effects further along the innovation chain.
Analysis of the effects of the R&D tax incentive further along the innovation chain: turnover, value added and labour productivity
The analysis applies the same difference-in-differences (DiD) approach and firm sample as in Growth Analysis (2022) and Stavlöt and Svensson (2025). The effect is identified by comparing firms below the cap with those that have reached it and thus no longer have an incentive to expand their R&D staff. The dataset covers the period 2009–2019 and combines information on employers’ social security contributions and R&D tax relief with firm-level data from Statistics Sweden (SCB) and data on direct enterprise support from government agencies. The outcome variables are turnover, value added and labour productivity, while the controls include firm size, age, solvency and capital intensity. The sample is restricted to established firms with continuous R&D activity to reduce the risk that results are driven by temporary or misreported claims.
The R&D tax incentive is associated with increased turnover
The results show clear and statistically significant effects on firms’ turnover. These effects appear about two years after the incentive was used, consistent with the time lag that can be expected between increased R&D investments and measurable economic outcomes. The magnitude of the effect is economically substantial, corresponding to an average increase of just over 20 per cent, or between SEK 0.4 and 1.1 million for the median firm. For value added, the estimated coefficients are positive but cannot be interpreted causally since the parallel-trend assumption cannot be confirmed. No statistically significant effects are found for labour productivity, and the assumption of parallel trends is not fulfilled. Overall, the findings suggest that the R&D tax incentive contributed to increased turnover, but it remains unclear whether this development reflects value-creating growth or rather an expansion in business activity without corresponding efficiency gains.
An alternative interpretation of the expected mechanism, where growth is driven by additional R&D investments, is that the observed turnover effects are partly explained by the liquidity released when the tax reduction lowers labour costs also for existing R&D staff. These freed-up resources can be used for expansion or recruitment in other parts of the firm. Although the additional liquidity is smaller among firms below the cap, its relative importance may be greater since these firms are typically smaller. Previous studies of other payroll tax reductions indicate that such scale effects may play an important role for smaller firms. At the same time, there is evidence that R&D staff has actually increased among firms below the cap, suggesting that the tax incentive effects have also contributed. It is therefore likely that both mechanisms operate simultaneously, and their relative importance should be explored in future research.
Growth effects confined to firms that also received direct subsidies
The heterogeneity analysis reveals no systematic differences between small and large firms, nor between firms with different levels of indebtedness. However, there are weak indications of stronger effects among more labour-intensive firms, consistent with the design of the R&D tax incentive. A more detailed analysis of the interaction with direct subsidies further shows that the turnover effects occur only among firms that also received project grants. This may suggest that the two support instruments act as complements, but it may also reflect that more R&D-intensive and resource-rich firms are more likely to apply for and obtain multiple forms of support. Such firms may have greater capacity to benefit from both the tax incentive effects and the financial resources arising from tax reductions and direct grants. These results should therefore be interpreted with caution, but they indicate that a combination of policy instruments may be an important condition for generating growth. The question of how different support schemes interact is important and warrants further analysis, although it is not explored in detail in this report.
Growth exceeding the R&D tax relief, but more knowledge is needed on its underlying drivers
Comparisons between the estimated turnover effects and the total value of the deductions show that the effects are large even relative to the governments forgone tax revenue. The findings indicate that the incentive may have been efficient, but uncertainty remains regarding how much of the observed turnover increase represents productivity-driven growth and how persistent the effects may be.
In summary, the study contributes to a more comprehensive understanding of the effects of the Swedish R&D tax incentive. The results indicate that the scheme has, on average, had substantial positive effects on firms’ turnover, although the evidence for productivity-driven growth is less clear. They also suggest that the tax incentive may be more effective when combined with direct support. For policy design, this implies that the interaction between different support measures deserves greater attention and that a coordinated strategy may yield stronger impacts than isolated instruments. Further research is needed to clarify the mechanisms driving the observed effects, the interaction between support schemes, and the long-term efficiency of the R&D tax incentive.
