Periodic Reporting for period 1 - DIV_INV (Knowledge Diversity Building by Inventors)
Okres sprawozdawczy: 2020-09-01 do 2022-08-31
The knowledge held by inventors and scientists is arguably the most important input to innovation. Yet, how knowledge built up by inventors affects the inventive process is scarcely understood. The main objective of the DIV_INV project is to collect evidence on how inventors' knowledge profiles relate to creative output, shedding light on the tension between the benefits of specialization and the costs of coordinating teamwork.
The project has 3 main scientific objectives. First, it seeks to understand to what extent diversity built up by teams of inventors can substitute for the diversity held by individual team members. Such substitutability is important to understand the potential friction brought about by the trend towards individual specialization and increased use of teamwork. If substitution is hard, this trend implies an important constraint on the production of knowledge, and therefore future productivity growth in the economy.
Second, the project asks whether policies that increase team breadth are likely to be effective in spurring more innovation. Such policies are commonplace to increase the efficiency of public funding. Most funding strategies assume that promoting multidisciplinary teams – that is, broader ones – will increase the innovation output generated by a given subsidy. The project seeks to provide causal evidence on the justification of this assumption and the policies it implies.
A third objective is to understand the extent to which monetary incentives effectively induce field-switching by inventors. This question is important to study demand-side policies' effectiveness in supporting innovation. Most innovation support schemes effectively increase the demand for innovation by increasing innovator profits. Recent examples include support schemes to induce innovations that mitigate climate change, for instance by developing renewable energy innovations. However, the supply of inventors in a given field conditions the effectiveness of such demand-side policies. Understanding how financial incentives affect field-switching is key to informing policymakers.
The project has 3 main scientific objectives. First, it seeks to understand to what extent diversity built up by teams of inventors can substitute for the diversity held by individual team members. Such substitutability is important to understand the potential friction brought about by the trend towards individual specialization and increased use of teamwork. If substitution is hard, this trend implies an important constraint on the production of knowledge, and therefore future productivity growth in the economy.
Second, the project asks whether policies that increase team breadth are likely to be effective in spurring more innovation. Such policies are commonplace to increase the efficiency of public funding. Most funding strategies assume that promoting multidisciplinary teams – that is, broader ones – will increase the innovation output generated by a given subsidy. The project seeks to provide causal evidence on the justification of this assumption and the policies it implies.
A third objective is to understand the extent to which monetary incentives effectively induce field-switching by inventors. This question is important to study demand-side policies' effectiveness in supporting innovation. Most innovation support schemes effectively increase the demand for innovation by increasing innovator profits. Recent examples include support schemes to induce innovations that mitigate climate change, for instance by developing renewable energy innovations. However, the supply of inventors in a given field conditions the effectiveness of such demand-side policies. Understanding how financial incentives affect field-switching is key to informing policymakers.
The first objective is to provide evidence on the substitutability of two types of knowledge diversity as input to the inventive process: (1) Diverse knowledge collected by individual inventors and (2) diverse knowledge collected through the combination of non-overlapping expertise by teams of inventors. The literature analysis showed two key problems the empirical design of the paper needed to solve. 1) Most contributions rely on cross-sectional analyses vulnerable to various endogeneity concerns. 2) No rigorous method exists to disentangle knowledge diversity at the level of the individual from knowledge diversity through team composition. The empirical design addresses the first issue by constructing a firm-level panel of inventive output and team diversity and a large set of controls to account for firm- and inventor-specific characteristics. Running regressions with multiple levels of fixed effects and various time-varying controls of inventor quality addresses omitted variable bias and can account (at least partly) for endogeneity issues. The second issue was solved by designing a new method to measure within-individual and across-individual knowledge diversity in teams of inventors. The results show that firms easily substitute both types of knowledge diversity.
The second objective is to collect causal evidence of the effect of increasing knowledge diversity on innovation output. The researcher identified a methodology to induce 'exogenous' variation in team breadth to address the issue that inventors select into teams based on unobservable project characteristics and individual traits. The design relies on the death of inventors as an exogenous shock to team composition in the future. The idea is as follows: if an inventor with overlapping knowledge with her teammates dies, future team breadth will not be affected because of the redundancy of knowledge. Instead, if an inventor with non-overlapping knowledge is lost, future team diversity will be impacted. Comparing the future innovation output of team members across these two cases helps to identify the causal effect of team breadth on innovation output. More traditional econometric analyses of team breadth and innovation quality are performed to corroborate the findings further.
The final objective seeks to understand the effectiveness of monetary incentives to make inventors build up expertise in a new technological domain. Addressing this question requires measuring when private returns to inventions in a given field increase and analyze how such an increase affects the probability of inventors outside the field engaging with that field. To estimate models of inventor supply in response to shocks to the private value of inventions in a field, we set up a panel data set of more than 800 technological fields. Measures of the private value of individual patented inventions from the literature are aggregated to this field level. At the same time, we collect the number of inventors active in each field to track changes to the supply of inventors. We estimate the effect of changing the private value of innovating in different fields on the supply of new inventors, controlling for technological opportunity in the area. The first analyses show that the inventor supply does not strongly react to changes in private value, indicating low inventor supply elasticities. We devised a causal design to verify our findings to address the concern that our controls do not pick up all relevant heterogeneity. We use three law changes that make relative patent protection across fields stronger/weaker. These law changes should be unrelated to technological opportunity or other unobserved confounders and therefore identify the causal effect of interest.
The primary mode of disseminating the results is through conference presentations and academic publications. Two papers are being prepared to be submitted to leading journals in the field (Strategic Management Journal and Management Science). The results related to the third objective will be summarized in an academic paper. Each publication will be posted in the portal. The researcher has presented at over ten international conferences and four seminars throughout the project.
The second objective is to collect causal evidence of the effect of increasing knowledge diversity on innovation output. The researcher identified a methodology to induce 'exogenous' variation in team breadth to address the issue that inventors select into teams based on unobservable project characteristics and individual traits. The design relies on the death of inventors as an exogenous shock to team composition in the future. The idea is as follows: if an inventor with overlapping knowledge with her teammates dies, future team breadth will not be affected because of the redundancy of knowledge. Instead, if an inventor with non-overlapping knowledge is lost, future team diversity will be impacted. Comparing the future innovation output of team members across these two cases helps to identify the causal effect of team breadth on innovation output. More traditional econometric analyses of team breadth and innovation quality are performed to corroborate the findings further.
The final objective seeks to understand the effectiveness of monetary incentives to make inventors build up expertise in a new technological domain. Addressing this question requires measuring when private returns to inventions in a given field increase and analyze how such an increase affects the probability of inventors outside the field engaging with that field. To estimate models of inventor supply in response to shocks to the private value of inventions in a field, we set up a panel data set of more than 800 technological fields. Measures of the private value of individual patented inventions from the literature are aggregated to this field level. At the same time, we collect the number of inventors active in each field to track changes to the supply of inventors. We estimate the effect of changing the private value of innovating in different fields on the supply of new inventors, controlling for technological opportunity in the area. The first analyses show that the inventor supply does not strongly react to changes in private value, indicating low inventor supply elasticities. We devised a causal design to verify our findings to address the concern that our controls do not pick up all relevant heterogeneity. We use three law changes that make relative patent protection across fields stronger/weaker. These law changes should be unrelated to technological opportunity or other unobserved confounders and therefore identify the causal effect of interest.
The primary mode of disseminating the results is through conference presentations and academic publications. Two papers are being prepared to be submitted to leading journals in the field (Strategic Management Journal and Management Science). The results related to the third objective will be summarized in an academic paper. Each publication will be posted in the portal. The researcher has presented at over ten international conferences and four seminars throughout the project.
The results from the project have clear policy implications. Most developed economies have subsidy schemes that promote within-researcher knowledge diversity or broader teams through multidisciplinary collaboration. The evidence base to justify knowledge diversity-building interventions has been thin. The research results suggest that the 'conventional wisdom' – more knowledge diversity is always better – might be misleading. Furthermore, the results bear relevance to a growing interest in industrial policy in Europe and beyond. The main focus of such an approach is to promote innovation in certain high-externality fields through demand-side incentives such as subsidies. Such instruments effectively increase the demand for innovation in a certain area by increasing the private returns to innovation. Given the supply-side frictions uncovered by this project, it is far from clear that such demand-side policies are optimal. Indeed, policies that target the supply of inventors to crucial fields, in the long run, seem necessary as short-run switching between areas seems unattractive to most inventors.