R\&D subsidies, income taxes, and growth through cycles (Q6063090)

Extending the work of \textit{K. Matsuyama} [J. Econ. Theory 100, No. 2, 220--234 (2001; Zbl 0992.91072)], the authors present models where, in line with the data analyzed for the US and 15 other OECD economies, household savings have negative correlation with growth while R\&D has positive correlation with growth. The endogenous growth-cycle model developed by the authors accounts for taxes on income and consumption, subsidies to investment and innovation, and R\&D labor and partial depreciation. Three main effects of income taxes and R\&D subsidies are found to account for the observed cyclical properties: \begin{itemize} \item Increasing R\&D subsidies raises monopolistic R\&D beyond its inefficient level, raising productivity by trading the intensive margin for the extensive margin due to diminishing marginal products, and thus is conducive to growth. \item R\&D-subsidy payments in periods with active R\&D increase household taxable income beyond output, increasing consumption and decreasing household saving rates, hindering future R\&D and growth. \item Anticipated R\&D subsidies raise household saving rates by raising returns to saving, promoting future growth. \end{itemize} The positive effects on R\&D and negative effects on saving of income tax financed R\&D subsidies in the Romer regime destabilize balanced growth and period-2 cycles. High R\&D subsidies and income taxes engender new patterns of fluctuations such as unstable period-4 cycles at plausible growth rates and plausible elasticity values of product substitution. Among the new patterns, unstable fluctuations with consecutive periods of horizontal R\&D prevail at realistic tax and subsidy rates in developed countries, and behave as if it were dependent on a random process.