We extend the basic Schumpeterian endogenous growth model by allowing incumbents to undertake innovations to improve their products, while entrants engage in more "radical" innovations to replace incumbents. Our model provides a tractable framework for the analysis of growth driven by both entry of new …rms and productivity improvements by continuing …rms. Unlike in the basic Schumpeterian models, subsidies to potential entrants might decrease economic growth because they discourage productivity improvements by incumbents in response to reduced entry, which may outweigh the positive e¤ect of greater creative destruction. As the model features entry of new …rms and expansion and exit of existing …rms, it also generates a non-degenerate equilibrium …rm size distribution. We show that, when there is also costly imitation preventing any sector from falling too far below the average, the stationary …rm size distribution is Pareto with an exponent approximately equal to one (the so-called "Zipf distribution").We thank Sam Kortum, Erzo Luttmer, Ariel Pakes, and seminar participants at MIT and Toulouse Network on Information Technology Conference at Seattle for useful comments. We are particularly grateful to Xavier Gabaix for numerous useful suggestions at the early stages of this project. Financial support from the Toulouse Network on Information Technology is gratefully acknowledged.

No abstract

We extend the basic Schumpeterian endogenous growth model by allowing incumbents to undertake innovations to improve their products, while entrants engage in more "radical" innovations to replace incumbents. Our model provides a tractable framework for the analysis of growth driven by both entry of new …rms and productivity improvements by continuing …rms. Unlike in the basic Schumpeterian models, subsidies to potential entrants might decrease economic growth because they discourage productivity improvements by incumbents in response to reduced entry, which may outweigh the positive e¤ect of greater creative destruction. As the model features entry of new …rms and expansion and exit of existing …rms, it also generates a non-degenerate equilibrium …rm size distribution. We show that, when there is also costly imitation preventing any sector from falling too far below the average, the stationary …rm size distribution is Pareto with an exponent approximately equal to one (the so-called "Zipf distribution").We thank Sam Kortum, Erzo Luttmer, Ariel Pakes, and seminar participants at MIT and Toulouse Network on Information Technology Conference at Seattle for useful comments. We are particularly grateful to Xavier Gabaix for numerous useful suggestions at the early stages of this project. Financial support from the Toulouse Network on Information Technology is gratefully acknowledged.

Under limited commitment that prevents agents from pledging their future non‐financial wealth, agents with incorrect beliefs always survive by holding on to their non‐financial wealth. Friedman's () market selection hypothesis suggests that their financial wealth trends towards zero in the long run. However, I present a dynamic general equilibrium model with incomplete markets due to collateral constraints and show that the hypothesis depends on the degree of market incompleteness. When markets are more incomplete, over‐optimistic agents not only survive but also prosper by speculation. Stricter margin requirements protect the wealth of the optimists and thereby increase asset price volatility.

No abstract