Matthew effects and R&D subsidues: knowledge cumulability in high-tech and low-tech industries

The paper explores the causes and effects of persistence in the discretionary allocation of public subsidies to R&D activities performed by private firms in high-tech and low-tech industries. It applies the crucial distinction between past dependent reputation-Matthew-effects and path dependent competence-Mattheweffects. The former qualifies the persistence in the discretionary allocation of public subsidies in terms of sheer information externalities exclusively based upon previous awards. The latter is identified by the role of the accumulation of competence stemming from past grants in current R&D activities. The paper articulates and tests the hypothesis that knowledge cumulability matters in assessing whether vicious or virtuous Matthew effects prevail. Competence-Matthew-effects are identified by the actual increase of total R&D activities of the recipients of public grants in the past. Virtuous Matthew effects are found in high-tech industries where learning, learning to learn and knowledge cumulability are higher. In traditional industries, vicious Matthew effects prevail for the lower levels of knowledge cumulability. Here reputation-Mattheweffects can lead to substitution of private funds with public ones. A rich and detailed empirical analysis including Transition Probability Matrices, probit regression and Propensity Score Matching on a database of around 700 Italian firms in the years 1998-2003, confirms the hypothesis and suggests that the selective use of discretionary allocation should be applied in high tech industries. The identification and appreciation of the key role of knowledge cumulability can become a major target for an effective innovation policy