192Our ﬁndings complement recent papers (e.g., Autio & Rannikko, 2016; Howell, 2017; Stevenson et al., 2021; Söderblom et al., 2015) that found signiﬁcant positive effects of government subsidies toward innovative SMEs. Their approach of using ﬁrms that applied for, but did not receive, funding as a control group has led to Do Targeted R&D Grants toward SMEs Increase Employment and Demand for High. . .

193varying outcomes. We cannot exclude that their results are due to an inherent selection bias in their analysis since the treatment group has been judged as more promising than the control group. This suggests that the treatment group would have performed better than the control group even without a grant.

The lack of signiﬁcant employment effects of the government support programs that we investigate is troublesome considering that policymakers often justify targeted R&D grants with the need to correct market failures and promote job growth. We believe that the lack of signiﬁcant results points toward the challenges involved in using targeted R&D grants as a way of promoting future growth among SMEs. Coad et al. (2014), for example, noted that it is very difﬁcult to point out which ﬁrms are going to be fast growers in the future, suggesting that government support programs are unlikely to target potential high-growth ﬁrms that would not grow without support. This is consistent with previous evidence that ﬁrm growth, to a large extent, can be considered random (Coad et al., 2013), thereby making it extremely hard for policymakers to determine what characterizes SMEs that need an R&D grant to promote job growth and demand for highly skilled labor.

Our ﬁndings may also reﬂect the heterogeneous nature of SMEs, even highly innovative ones. Reﬂecting the variety of innovative ﬁrms may begin to lessen the randomness of the next stages of development, by including truly new ventures with their initial product offering, developing, or commercializing a new-to-the-world technology vs. leveraging innovation from elsewhere, younger ventures that are highly technical but which may not rely on traditional R&D functions, or even more established small ﬁrms looking to expand. One notion advanced by Mason and Brown (2013) is to focus on outcomes that help support retaining winners rather than simply picking winners. This approach would at a minimum remove some sources of variance among ﬁrms applying for these types of growth grants.

As an alternate explanation, it may be that the highly inﬂuential interviews and expert evaluations of those ﬁrms under consideration are ineffective. There is currently a dearth of empirical evidence that scrutinizes the questions asked, of whom they are asked, and how the answers are analyzed as part of the application process—or what objective metrics they employ to rate the attractiveness of these potential grant recipients. Even when professional investors have tremendous difﬁculties in predicting the future outcomes of high-potential but risky ventures, policymakers maybe even less equipped to make these evaluations or provide monitoring of the funding over time (Lerner, 2009). The opacity of this evaluation process and the inconclusive results of their beneﬁt brings into question whether and how robust and objective decision rubrics can effectively be employed.

More broadly, the absence of positive results in our study brings into question whether government support programs toward SMEs can be justiﬁed, given that they are associated with high administrative costs, increased incentives for rent-seeking behavior among entrepreneurs, and crowding-out effects on alternative investments that could be more beneﬁcial for society (Bradley et al., 2021). The incentive and ability for researchers to publish results that are statistically signiﬁcant (Møen & Thorsen, 2017) might also have led to an overconﬁdence in policymakers’ abilities S.-O. Daunfeldt et al.

194to inﬂuence the future growth and human capital of SMEs, when the non-ﬁndings of policy effects are rarely published. Our study does not come without limitations, however. Even though we use a matching method that is at the research frontier, our results might still be biased if unmeasured variables are correlated with job growth and the likelihood of receiving a targeted R&D grant. We believe that the approach used by Söderblom et al. (2015) can provide more reliable estimates if the policymakers in the last stage of the decision-making process would randomize which ﬁrms receive an R&D grant. The advantage of using a randomized ﬁeld experiment is that the outcome variable cannot affect the probability of receiving an intervention, which means that we know that it is not the intervention that affects the outcome variable. Randomization also implies that there is no systematic connection between the probability of belonging to the intervention group and observable and non-observable factors (Burtless, 1995). Randomized control trials (RCT) have recently been used in the UK to evaluate the effects of targeted R&D grants (Bakhshi et al., 2015; Roper, 2020), and McKenzie (2017) provides fascinating evidence from an RCT in Nigeria on the effects of public grants following a national business program competition. We believe that more such studies are needed to provide more robust evidence on the effectiveness of government support programs, although we recognize that introducing randomization may be a challenge for policymakers to justify. But it may also potentially remove concerns regarding implicit bias (or crony capitalism at worst; Klein et al., 2021) from selecting from among a group of SMEs that otherwise meet or exceed the criteria for a grant. Another fruitful area for future research is to more closely evaluate whether the effects of the grants are related to underlying unobservable or difﬁcult-to-quantify factors, such as differences in how well companies are integrated into local business conditions, or the presence of positive spillovers from other companies. These factors are found to be important for growth among innovative ﬁrms but create potential challenges to identify and categorize a priori. More research is also needed on whether certain types of targeted R&D grants are more effective than others. Certainly, innovative activities among SMEs come in many shapes and sizes (McKelvie et al., 2017), where many of the most impactful aspects that lead to growth do not appear as formal R&D activity. As such, a heterogeneity analysis could deepen the understanding of the conditions under which the opportunities for positive effects of government support programs are greatest across different aspects of innovative activities beyond R&D. Instead of focusing on a small group of growth-oriented R&D-intensive SMEs, it may also be more important to focus research and policy measures on what is needed to stimulate growth among SMEs that do not grow, or at best grow marginally. As noted by Bornhäll et al. (2015), the existence of growth barriers is likely to prevent these ﬁrms from growing, while potential high-growth ﬁrms might grow despite the existence of such growth barriers. General policy measures aimed at low-growth SMEs (e.g., simpliﬁcation of rules, reduced labor costs, more liberal employment protection legislation, etc.) can thus be more effective in promoting job growth than Do Targeted R&D Grants toward SMEs Increase Employment and Demand for High. . .

195targeted R&D grants toward SMEs that are considered potential high-growth ﬁrms. This corresponds to the analogy in Coad et al. (2014, p. 92), in which tourists on safaris are focused on beautiful gazelles but fail to see the importance of the dung beetle in maintaining the health of the ecosystem.

Note also that we have been investigating the effects of two Swedish R&D grants that are administrated by Vinnova and targeted toward growth-oriented SMEs. The lack of signiﬁcant effects does not mean that R&D grants never work. As noted in our literature review, there is substantial evidence that grants can indeed be helpful for SMEs when looking at different outcome variables. Other types of programs can also be successful in promoting employment growth, while similar R&D grants might be efﬁcient under other institutional contexts (e.g., in other countries). The unique nature of growth-oriented ventures in Sweden has been noted in the literature (McKelvie et al., 2021). Another possible interpretation of our non-signiﬁcant results is that the R&D grants under study usually work, but that the government agency administering the programs is not designing or executing the programs adequately. The R&D grants might, for example, work better if they were given to larger ﬁrms or a smaller set of higher-quality applicants. Our study does not currently investigate these alternate models, but we do encourage others to take up this task. We have furthermore focused on the employment effects of targeted R&D grants; it is possible that such programs have a higher impact on other outcome variables (see e.g., Howell, 2017).6 The external validity of our ﬁndings is thus low, and interpretations in other contexts should be made with care. This highlights the importance of gathering more robust evidence on the effects of government support programs that are targeted toward SMEs with growth potential. While we fully accept that the development of growth-oriented SMEs is important to the economy, we also recognize that more transparent and methodologically sophisticated tools are needed to more fully evaluate the effectiveness of current practices, such as R&D grant programs. Our intent with this study is to illustrate two programs that are well-intended but that do not seem to have the desired impact, and to offer thoughts on the conditions through which we as scholars can better make these determinations of effectiveness. In doing so, we hope to contribute to a more robust and systematic understanding of how government policies further—or fail to further—entrepreneurship.

Acknowledgments We gratefully acknowledge funding from the Swedish Competition Authority (grant number 376/2016).

6Note that the authors have investigated the effects of these R&D grants on sales and labor productivity in a Swedish report for the Swedish Government Agency for Growth Policy Analysis (Daunfeldt et al., 2014). Their results indicate no statistically signiﬁcant effects on labor productivity, while the government support programs increased sales by 20% during the ﬁrst post-support year. The latter effect is, however, limited to ﬁrms with at most six employees. No positive and statistically signiﬁcant effect is found for larger ﬁrms.

196S.-O. Daunfeldt et al.

References

Angrist, J. D., & Pischke, J. S. (2008). Mostly harmless econometrics. Princeton University Press. Autio, E., & Rannikko, H. (2016). Retaining winners: Can policy boost high-growth entrepreneurship? Research Policy, 45(1), 42–55. Bakhshi, H., Edwards, J. S., Roper, S., Scully, J., Shaw, D., Morley, L., & Rathbone, N. (2015). Assessing an experimental approach to industrial policy evaluation: Applying RCT+ to the case of creative credits. Research Policy, 44(8), 1462–1472. Baumol, W. J. (1990). Entrepreneurship: productive, unproductive and destructive. Journal of Political Economy, 98(5), 893–921. Becker, L. (2015). Effectiveness of public innovation support in Europe. Does public support foster turnover. In Employment and labour productivity? Discussion papers, Center for European Governance and Economic Development Research (Vol. No. 236). University of Goettingen. Blackwell, M., Iacus, S. M., King, G., & Porro, G. (2009). Cem: Coarsened exact matching Stata. Stata Journal, 9(4), 524–546. Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1), 115–143. Bornhäll, A., Daunfeldt, S. O., & Rudholm, N. (2015). Sleeping gazelles: The unseen job creators? In A. C. Corbett, J. A. Katz, & A. McKelvie (Eds.), Entrepreneurial growth: Individual, ﬁrm, and region. Emerald Group Publishing Limited. Bradley, S. W., Kim, P. H., Klein, P. G., McMullen, J. S., & Wennberg, K. (2021). Policy for innovative entrepreneurship: Institutions, interventions, and societal challenges. Strategic Entrepreneurship Journal, 15(2), 167–184. Bronzini, R., & Iachini, E. (2014). Are incentives for R&D effective? Evidence from a regression discontinuity approach. American Economic Journal: Economic Policy, 6(4), 100–134. Burtless, G. (1995). The case for randomized ﬁeld trials in economic and policy research. Journal of Economic Perspectives, 9(2), 63–84. Cantner, U., & Kösters, S. (2015). Public R&D support for newly founded ﬁrms–effects on patent activity and employment growth. Journal of Innovation Economics Management, 1(16), 7–37. Coad, A., Frankish, J., Roberts, R. G., & Storey, D. J. (2013). Growth paths and survival chances: An application of Gambler's ruin theory. Journal of Business Venturing, 28(5), 615–632. Coad, A., Daunfeldt, S. O., Hölzl, W., Johansson, D., & Nightingale, P. (2014). High-growth ﬁrms: Introduction to the special section. Industrial and Corporate Change, 23(1), 91–112. Daunfeldt, S. O., Gustavsson Tingvall, P., & Halvarsson, D. (2014). Företagsstöd till innovativa små och medelstora företag: En kontrafaktisk effektutvärdering, report 2014:26, the Swedish Government Agency for Growth Policy Analysis. Dvouletý, O., Srhoj, S., & Pantea, S. (2021). Public SME grants and ﬁrm performance in European Union: A systematic review of empirical evidence. Small Business Economics, 57(1), 243–263. Edler, J., Cunningham, P., Gök, A., & Shapira, P. (2013). Impacts of innovation policy: Synthesis and conclusions–compendium of evidence on the effectiveness of innovation policy intervention project. Nesta, 2013, June. UK. Hansson, P. (2000). Relative demand for skills in Swedish manufacturing: Technology or trade? Review of International Economics, 8(3), 533–555. Hayakawa, K. (2009). First difference or forward orthogonal deviation-which transformation should be used in dynamic panel data models: A simulation study. Economics Bulletin, 29(3), 2008–2017. Henrekson, M., & Johansson, D. (2010). Gazelles as job creators: A survey and interpretation of the evidence. Small Business Economics, 35(2), 227–244. Hijzen, A., & Swaim, P. (2008). Offshoring, Labour Market Institutions and the Elasticity of Labour Demand.The University of Nottingham, Research paper series, No. 2008/05. Himmelberg, C. P., & Petersen, B. C. (1994). R&D and internal ﬁnance: A panel study of small ﬁrms in high-tech industries. The Review of Economics and Statistics, 76(1), 38–51.