Investing in R&D: small- and medium-sized enterprise financing preferences

Abstract

This work adds to our understanding of financing decisions among owners of small- and medium-sized enterprises (SMEs), with particular reference to SMEs that conduct research and development (R&D). The work examines, conceptually and empirically, the forms of financing that are preferred by primary owners of SMEs that invest in R&D relative to preferences of owners of SMEs that do not conduct R&D. The work develops a conceptual rationale as to why SMEs engaged in R&D might hold particular preferences with respect to preferred sources of financial capital. Empirical analysis draws on large-scale survey data of actual applications for financing (as opposed to financing received) as reported by business owners and controls for systemic factors that include firm size and sector. Findings include that financing preferences do not follow a ‘one-size-fits-all’ prescription. Rather, preferences vary according to growth goals, the nature of ownership, age of firm, firm size and sector; however, it was also clear that firms that invest in R&D are much more likely to seek equity financing than otherwise comparable firms that do not invest in R&D.

Keywords:

• financing preferences
• R&D
• innovation
• SME

Notes

 1. The ‘pecking order’ theory of financing choices (Myers and Majluf 1984) suggests that internal capital is preferred to external debt, which, in turn, is preferred to external equity. However, firms that invest in R&D are arguably relatively informationally-opaque (Binks and Ennew 1996; Brierley 2001) and ought, therefore, to be more likely to seek non-debt external financing (MacIntosh 1994; Aboody and Lev 2000; Kortum and Lerner 2000).

 2. For example, early-stage firms are more likely to seek external financing; in addition loan applications from early-stage firms are rejected significantly and materially more often than those from more mature enterprises (www.sme-fdi.ic.gc.ca).

 3. The pecking order theory posits that businesses' first preference is to raise finance from internal sources. If external sources are necessary, debt is preferred over external equity because of the adverse signaling associated with a secondary offering of common shares.

 4. However, these findings are based on the types of financing actually used by firms: certain types of firms (new firms, technology-based firms, high-growth firms, etc.) typically face high rates of turndowns for financing requests, so usage may differ from preferences.

 5. There is an extensive literature regarding financing gaps and capital rationing. In essence, the theory of financing gaps is vested in the assumption that information asymmetry limits lenders' and investors' ability to assess – and price to – risk. In practice, both Cressy (2002) and Parker (2002) have concluded, in their respective reviews of the literature on financing gaps, that empirical evidence about the existence and materiality of gaps is ambiguous.

 6. For example, the Government of Canada has defined KBIs, or knowledge-based industries, according to a specific set of NAIC codes that correspond to the technological content of the goods or services sold by the firm. See: Industry Canada. 2000. Comparison and reconciliation of SIC and NAICS industry codes used to define knowledge-based industries (KBIs). http://www.sme-fdi.gc.ca/eic/site/sme_fdi-prf_pme.nsf/eng/h_01299.html (accessed December 29, 2010).

 7. However, because the questionnaire failed to solicit information on the actual value of ‘total investment expenditure’ the data only allow an interpretation as to whether or not there was any investment in R&D.

 8. There is a debate in the literature with respect to the relative advantages of logit or probit models. According to Hahn and Soyer (2005, 1), ‘Current opinion regarding the selection of link function in binary response models is that the probit and logit links give essentially similar results. … We find clear evidence that … the logit link provides better fit in the presence of extreme independent variable levels. Conversely, model fit in random effects models with moderate size data sets is improved generally by selecting the probit link.’

 9. These values are not unusual for cross-sectional data.

10.  http://www.sme-fdi.gc.ca/eic/site/sme_fdi-prf_pme.nsf/eng/02152.html, Table 18.