Governments subsidize research and development (R&D) through a mix of interdependent mechanisms, but the implications of subsidy interdependencies are not well understood. In this paper, I implement two quasi-experimental research designs to evaluate how the layering of innovation subsidies impacts firm R&D activity. I use funding rules and policy changes in the United Kingdom that generate exogenous variation in the cost of investing in R&D and find that direct grants and tax credits for R&D are complements for small firms but substitutes for larger firms. The effects are large. An increase in the tax credit rate enhances the effect of grant funding on small firms’ R&D expenditures so much that R&D expenditures more than double. For larger firms, higher tax credit rates cut the positive effect of grant funding in half. I also show that subsidy interactions influence the types of innovation efforts that emerge: with increases in both subsidy types, small firms steer efforts increasingly towards developing new goods as opposed to improving existing ones. I explore the underlying mechanisms and provide evidence that the complementarity for small firms is a result of easing financial constraints. Substitution by larger firms is consistent with public resources subsidizing infra-marginal R&D expenditures. Some alternative explanations can be ruled out, such as expenditure relabelling and inelastic inputs. Accounting for subsidy interactions is important for optimal innovation policy design and doing so could substantially improve the effectiveness of public spending on R&D.
Prosumer Consumption Substitution and Disentangling Salience from Uncertainty (with Eoghan McKenna) [draft undergoing revisions]
We study prosumers to disentangle whether information affects behavior by reducing uncertainty or increasing salience. Using high-frequency data on solar PV households, we show that consumption substitution of dirty electricity for clean electricity increases in solar endowment levels but decreases in its variability. Real-time information offsets the uncertainty effect significantly but does not change the level effect. We use econometric and machine learning methods to derive consumption substitution counterfactuals and find that information increases consumer surplus substantially. The methods can be applied in other settings with misoptimizing consumers when within-subject treatment variation is unavailable.
Research in Progress
“The Impact of Environmental Policy on Jobs and Firm Performance in China” (with Yangsiyu Lu) [draft coming soon, extended abstract can be found here]
“Innovation and Entrepreneurship in the Energy Sector” (with David Popp, Ivan Hascic, and Nick Johnstone)
“Some Causal Effects of Innovation and Environmental Policy Interactions” (with Suganda Srivastav)
“Do Firms Innovate More with Subsidies? Evidence from Electric Vehicle Manufacturing in China” (with Yangsiyu Lu)
“R&D Subsidies and Directed Technological Change”
Unpublished Working Papers
“Steering the Climate System: An Extended Comment” (with L. Mattauch, R. Millar, F. van der Ploeg, A. Rezai, A. Schultes, F. Venmans, N. Bauer, S. Dietz, O. Edenhofer, N. Farrell, C. Hepburn, G. Luderer, F. Spuler, N. Stern, and A. Teytelboym)