Public Funding and Innovation Strategies. Evidence from Italian SMEs

Abstract

Using a dataset which combines Community Innovation Survey (CIS) and accounting information on Italian manufacturing firms, we adopt a two step econometric procedure to investigate whether the receipt of public funding determines firms’ innovation strategy selection. In the first step an innovation selection equation is estimated, in the second we adopt a bivariate probit model. The main finding is that public funding influences whether firms select the make, the buy or the make&buy strategy, favoring the latter. The composite strategy is the one linked to the build up of absorptive capacity which, according to the empirical evidence, turns out to be associated to a better innovative performance.

Keywords:

• Public funding
• innovation strategies
• CIS survey

JEL CLASSIFICATIONS:

• G32
• O31
• D21

Acknowledgements

The authors are grateful to the referees and to all participants in the RSA European Conference 2015 and ENEF 2015 Conference for their valuable comments and suggestions. This paper is a result of a joint research project between the Italian National Institute of Statistics (ISTAT, Regional Office for Lombardy) and Università Cattolica del Sacro Cuore (UCSC) entitled: ‘Social Capital, Innovation and Finance: empirical evidences on the manufacturing sector in Italy and in the Lombardy region’. All remaining errors are our own.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 The Barcelona European Council of March 2002, announced targets for rising R&D investment to 3% of GDP by 2010. The target of 3% was confirmed in 2010.

2 OECD (2006)

3 See for example Heim (2016).

4 See Zúñiga-Vicente et al. (2014), Hall et al. (2015) and the literature therein.

5 The Italian experience can be actually seen as an amplified version of the (average) European experience.

6 Using the data collected by the Bank for International Settlements and publicly available on their website, it is possible to compute the ratio between bank credit and total credit to the private nonfinancial sector which is typically used as a proxy of financial development. It turns out that this index was, on average, 72% between 2012 and 2016 in Italy, much higher than in the USA, France and the UK (respectively, 33%, 51%, and 53%) and similar to that in Germany (73%) where banks have traditionally played an important role in financing the economy. See https://www.bis.org

7 A firm that chooses to buy technology in the form of technologically advanced machinery, for example, can use the asset as collateral and collateral assets appear to reduce agency costs of debt.

8 See Laursen and Salter (2014) and Cassiman and Veugelers (2002).

9 The Micro_Manu.Istat 2000–2010 database is a result of collaboration between the Italian National Institute of Statistics (Istat, Regional Office in Lombardy) and Università Cattolica del Sacro Cuore (UCSC). More information may be found in ISTAT and Università Cattolica del Sacro Cuore, 2014.

10 This is a firm level survey, compiled every 4 years in all EU member states and some non-EU countries

11 The Italian Business Register, named Statistical Register of Active Enterprises (ASIA), has been developed in order to respect the European requirements regarding the realization of business registers based on administrative data. Business registers were necessary to the yearly production of harmonized official statistics for the whole population of non-agricultural enterprises, given that Censuses on the industrial system are normally taken every ten years. In 2008, in order to ensure a harmonised framework of the business registers, it was considered appropriate to adopt a new Regulation (EC N.177/2008 of the European Parliament and of the Council of 20 February 2008 establishing a common framework for business registers for statistical purposes and repealing EEC N.2186/93).

12 It should be noted that the ‘sample reduction’ resulting from the merging of CIS and balance-sheet data leads to a final sample biased in favor of larger and more innovative firms. Comparing our sample (3717 firms) with the universe of firms in CIS7 (18152 firms), we note an average firm size, in logarithmic terms, of 4.092 (60) employees versus 3.452 (32) in the CIS7 and a percentage of innovative firms of 58% versus 35% of CIS7.

13 However, in order to test whether our results are influenced by the presence of EU funding which, differently from other sources, promotes cooperative R&D and provides monitoring and screening activities, we conduct the same analysis excluding EU funds and obtain the same outcome. Results are available upon request.

14 Note that these three variables are dummy variables. In particular make is a dummy variable which assumes value 1 in the case the business engaged only in internal R&D and 0 otherwise; buy is a dummy which assumes value 1 if the firm conducted only extra muros R&D activities and 0 otherwise; finally make&buy is a dummy equal to 1 if the firm pursued a make&buy strategy during the period of analysis.

15 We have also tried a multinomial version of this model, which rather complicates the estimation procedure, but allows to estimate jointly over the whole sample of innovative firms. We have found very similar results, which are available upon request.

16 We consider the following Italian regions: North East, North West, Center and South. We consider the technological sectors according to the Pavitt classification divided into high, medium-high, medium-low and low.

17 Past innovative performance might also be a relevant variable in this context. We use a balanced dataset linking two sequential CIS waves to account for past innovative performance measured as R&D expenditure per employee. The impact of Public Funding on the strategies is confirmed also in this sub sample of 806 firms. Results are available upon request.

18 Given that the main source of Public Funding in our sample is mainly coming from National and Regional sources and not from EC Programmes, we presumably expect that cooperation variables influence the probability of receiving Public Funding and not viceversa, given that only European funding has the aim of promoting exclusively cooperative R&D.

19 We have been able to confirm the main message of the paper also performing a robustness check using a different methodology which does not rely on the validity of exclusion restrictions, i.e propensity matching. Results are available upon request.

20 Note that formal definitions of all the variables are provided in Appendix Table A1.