Non-Balanced Endogenous Growth and Structural Change: When Romer Meets Kaldor and Kuznets

We propose a model of non-balanced endogenous growth in which the final good, which can be either consumed or used as capital, is produced using two intermediate inputs, one being "knowledge-intensive". Agents working in the knowledge-intensive sector need to accumulate technological knowledge and thus have to decide how to split their individual unit of time between accumulation of technological knowledge (research) and work. Agents working in the second sector do not need to accumulate knowledge and thus devote all their individual unit of time to work. Individual knowledge therefore becomes a labor-augmenting factor, and knowledge accumulation leads to an unbounded increase in TFP in the knowledge-intensive sector, and thus to endogenous capital deepening. The asymmetry in the growth rates of TFP leads to non-balanced growth. Labor (number of workers) reallocations across sectors occur, leading to a greater increase in output for the knowledge-intensive sector. We show that non-balanced growth is consistent with Kaldor facts, as the asymptotic equilibrium is above all characterized by a constant interest rate and capital share in national income. However, the economy follows a growth path converging to a particular level of wealth that depends on the initial price of capital and knowledge. As a consequence, countries with the same fundamentals but lower initial wealth will be characterized by lower asymptotic wealth. We therefore extend the Lucas [19] finding and prove the existence of non-convergence across countries in a framework with structural change.