This dissertation studies various aspects of firm dynamics, and its relation to innovation and economic growth. Chapter 1 studies how incorporation, which provides limited liability protection to firm owners, affects firm behavior and economic growth. I propose an endogenous growth model of heterogeneous firms, where firms spend resources to improve their productivity and choose whether to incorporate or not. The model underlines incentive and selection effects of incorporation that generate the observed differences between incorporated and unincorporated firms. I calibrate the model to the Danish firm-level data to study the importance of these two effects quantitatively and draw several policy conclusions. Chapter 2 (joint with Ufuk Akcigit and Michael Peters) studies managerial delegation and its importance on the process of firm growth and aggregate productivity. We construct a model of firm dynamics where entrepreneurs have a fixed time endowment to run daily operations and need to hire outside managers as they grow. We calibrate the model to plant-level data from the US and India and quantify the importance of frictions in the managerial delegation to explain the differences in the firm dynamics and aggregate productivity between two countries. Chapter 3 (joint with Daron Acemoglu, Ufuk Akcigit, Nicholas Bloom and William Kerr) develops a model of endogenous reallocation and innovation with heterogeneous firms. Our main focus is on the reallocation (and misallocation) of R&D inputs, which emphasizes that misallocation may affect equilibrium growth as well. We estimate the model by using US Census Bureau micro data to study the effects of various counterfactual policies and gain insights about whether substantial improvements in economic growth and welfare are possible. Our results highlight the potential pitfalls of industrial policies supporting incumbents.