R&D and Productivity in OECD Firms and Industries: A Hierarchical Meta-Regression Analysis

Abstract

Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. AbstractThe relationship between R&D investment and firm/industry productivity has been investigated widely following seminal contributions by Zvi Griliches and others from late 1970s onwards. We aim to provide a systematic synthesis of the evidence, using 1,253 estimates from 65 primary studies that adopt the so-called primal approach. In line with prior reviews, we report that the average elasticity and rate-of-return estimates are positive. In contrast to prior reviews, however, we report that: (i) the estimates are smaller and more heterogeneous than what has been reported before; (ii) residual heterogeneity remains high among firm-level estimates even after controlling for moderating factors; (iii) firm-level rates of return and within-industry social returns to R&D are small and do not differ significantly despite theoretical predictions of higher social returns; and (iv) the informational content of both elasticity and rate-of-return estimates needs to be interpreted cautiously. We conclude by highlighting the implications of these findings for future research and evidence-based policy. AbstractThe relationship between R&D investment and firm/industry productivity has been investigated widely following seminal contributions by Zvi Griliches and others from late 1970s onwards. We aim to provide a systematic synthesis of the evidence, using 1,253 estimates from 65 primary studies that adopt the so-called primal approach. In line with prior reviews, we report that the average elasticity and rate-of-return estimates are positive. In contrast to prior reviews, however, we report that: (i) the estimates are smaller and more heterogeneous than what has been reported before; (ii) residual heterogeneity remains high among firm-level estimates even after controlling for moderating factors; (iii) firm-level rates of return and within-industry social returns to R&D are small and do not differ significantly despite theoretical predictions of higher social returns; and (iv) the informational content of both elasticity and rate-of-return estimates needs to be interpreted cautiously. We conclude by highlighting the implications of these findings for future research and evidence-based policy.

Studying abroad and earnings: A meta‐analysis

“…In the first model (whose results are presented in Column 1 of Table 6) we include all the potential explanatory variables 21 . Next, we follow a general‐to‐specific approach that consists in dropping the moderator variables with the largest p ‐value one at the time until the coefficients on the remaining variables are statistically significant (Ugur et al., 2016). As shown in Column 2 of Table 6, the general‐to‐specific process yielded a regression with four independent variables in addition to the constant (included by default).…”

Does intellectual property protection deliver economic benefits? A multi‐outcome meta‐regression analysis of the evidence

“…Moreover, most meta‐analysis studies focusing on a single outcome either overlook the issue of within‐study dependence or take account of the latter through clustered standard errors (e.g., Alptekin & Levine, 2012). Although some meta‐analysis studies take account of within‐study dependence and between‐study heterogeneity through multi‐level modeling (e.g., Awaworyi Churchill et al., 2017a; Ugur et al., 2018; Ugur et al., 2020; Ugur et al., 2016), the two‐level meta‐regression models in these studies are inadequate for synthesizing the evidence on the economic benefits of IP protection, where we have multiple effect‐size estimates for multiple outcomes that are theoretically related.…”