Unbundling innovation strategies: firms’ technological choices and complementarities-in-performance of innovation inputs and outputs across developing countries

Abstract

Innovation is considered a multi-level construct and it takes the form of (i) inputs and (ii) outputs. Based on this, the thesis considers three inter-related questions in the form of three essays: (1) What factors (firm, industry, institutional) affect innovation input choices and their combinations at the firm level? (2) How do complementarities in innovation inputs impact firm performance? and (3) How do complementarities in innovation outputs affect firm performance? The first essay explores the relationship between innovation input choices and their determinants based on firm, industry, and institutional factors. The strategic choice of innovation inputs by firms reflects the conditions under which make (R&D) and buy (purchase of technology) alternatives are complements or substitutes. If firms choose to combine R&D and purchased technology, the two are complements, but if the opposite happens, they are perceived as substitutes. We take a step back and understand which elements affect these innovation input choices by using factors rooted in RBV traditions and the TCE framework both individually and in combination to analyze their impact on such innovation inputs. Using the World Bank Enterprise Survey data of multiple developing and developed countries with Multinomial Logistic Regression, we explore how heterogeneity in firm characteristics, industry structures, and institutional regimes, along with their interactions, impact firms’ innovation input choices. We find, among other results that large firm size, availability of credit, and combining favourable regulatory environment with large firm size increases the likelihood of firms choosing to combine innovation inputs.The second essay investigates complementarities-in-performance of innovation inputs This examination is necessitated as no study explicitly looks at complementarities-in-performance of innovation inputs since all inputs do not result in measurable outputs, and similarly, not all innovation outputs can be traced back to such inputs. We conduct analyses with both the OLS and the Quantile Regression methods along with supermodularity. Our main finding is that low-performing firms tend to substitute R&D with External Technology Licensing rather than treat them as complements when analyzed through supermodularity. The third question extends the literature on complementarities-in-performance of innovation outputs and examines how investment in new forms of complementary assets through marketing innovation combined with product and process innovation impact firm performance using both OLS and Quantile Regression. Their complementarity is confirmed through supermodularity. Here too, we find that low-performing firms tend to treat the three forms of innovation as substitutes and not as complements.