This paper by Paul M. Romer explores the concept of growth driven by increasing returns due to specialization. It examines two dynamic models with specialized inputs and variety in production, demonstrating how these models can support continuous growth through monopolistically competitive equilibria. The paper builds on the idea that specialization can lead to increasing returns, a concept that has been recognized in economics for a long time. Romer focuses on the role of specialization rather than the accumulation of knowledge or externalities from knowledge spillovers, which are more commonly studied in models of increasing returns. The static models presented in the paper use a production function that captures the utility of variety, where output increases with the range of specialized inputs used. The models assume a cost function with a U-shaped average cost curve, which limits the degree of specialization. The paper derives equilibrium conditions and calculates the quantities of intermediate inputs and the price of the primary resource. It shows that the decentralized equilibrium can be suboptimal compared to a social planner's choice, leading to undervaluation of the primary resource. A dynamic model is then constructed by allowing for the accumulation of the primary resource, interpreted as durable capital. This model shows that the economy can exhibit unbounded growth, with consumption and capital stock growing at a constant rate. The dynamic equilibrium is suboptimal, and the paper suggests that interventions such as subsidies to savings could improve efficiency. The analysis highlights the presence of a "positive externality" associated with specialization, similar to Marshall's concept of an "external economy." The paper concludes by discussing the implications of these models for understanding economic growth and the role of specialization in driving long-term development.This paper by Paul M. Romer explores the concept of growth driven by increasing returns due to specialization. It examines two dynamic models with specialized inputs and variety in production, demonstrating how these models can support continuous growth through monopolistically competitive equilibria. The paper builds on the idea that specialization can lead to increasing returns, a concept that has been recognized in economics for a long time. Romer focuses on the role of specialization rather than the accumulation of knowledge or externalities from knowledge spillovers, which are more commonly studied in models of increasing returns. The static models presented in the paper use a production function that captures the utility of variety, where output increases with the range of specialized inputs used. The models assume a cost function with a U-shaped average cost curve, which limits the degree of specialization. The paper derives equilibrium conditions and calculates the quantities of intermediate inputs and the price of the primary resource. It shows that the decentralized equilibrium can be suboptimal compared to a social planner's choice, leading to undervaluation of the primary resource. A dynamic model is then constructed by allowing for the accumulation of the primary resource, interpreted as durable capital. This model shows that the economy can exhibit unbounded growth, with consumption and capital stock growing at a constant rate. The dynamic equilibrium is suboptimal, and the paper suggests that interventions such as subsidies to savings could improve efficiency. The analysis highlights the presence of a "positive externality" associated with specialization, similar to Marshall's concept of an "external economy." The paper concludes by discussing the implications of these models for understanding economic growth and the role of specialization in driving long-term development.