“Across the board, [the study’s] results imply that the additionality [or desired public policy benefit] of direct support might be on average slightly higher than for tax incentives. This finding, consistent with the literature, can be explained by the fact that most R&D tax incentives are provided on a non-discretionary basis.” [Note: later in the report, the authors point out the difference is not statistically significantly higher.]
In other words, not all R&D may be created equal with regard to public value or use. As a result, the cost to the public in forgone revenue to support other necessary public sector expenses that results from a company taking the R&D tax credit may not be necessarily commensurate with the company’s R&D expenditures. The authors suggest, but are not prepared to say this particular study provides the assessment required to recommend, that R&D tax credits, if used, may be most effective when applied in a non-discretionary fashion to those R&D expenditures which are expected to result in the greatest additionality.
Direct subsidy programs to support business R&D may not be without their own complications, implementation challenges and drawbacks, the authors caution. Higher administrative costs for government and higher compliance costs for firms are two examples. As a result, the authors offer a nuanced conclusion suggesting, “An optimal policy mix is likely to require a combination of discretionary and non-discretionary support elements.”
Perhaps of more value for policy makers interested in the efficacy of policy interventions to encourage business R&D investment is the opportunity to work directly with the publicly accessible OECD R&D Tax Incentives Database on which the paper is based. Users will be able to access first-ever, consistent time-series estimates of the cost of government tax relief for business R&D investment (foregone tax revenue and refunded amounts) for 2000 through 2016. Additionally, the database contains the new B-Index, presenting implied marginal R&D tax subsidy rates and profitability estimates for both SMEs and large firms from 2000 through 2018.