These essays attempt to explore how technological change, technology diffusion and economic distortions shape the aggregate economy. The first chapter empirically documents that wage inequality within the group of skilled workers in the U.S. has significantly widened since 2000 and that the changing trend of wage inequality was entirely driven by the non-routine analytic occupation. The model I build demonstrates that the task allocation induced by investment- specific technical change can widen the within-group wage inequality because of the “composition effect”. The quantitative results provide a well-matched timing and magnitude of the non-linear expansion path in wage inequality that is observed in the data. In chapter two I explore the role human capital plays in the convergence of Asian growth miracles. I incorporate the idea that education could facilitate technology diffusion into a growth framework by developing a model of human capital investment, adding a role for human capital in the convergence of productivities towards the technology frontier. I then calibrate my model to the South Korea between 1960 and 2019. My model can remarkably match the ‘S Shaped’ convergence trajectory in South Korea well. More importantly, the quantitative exercises demonstrate that a significant extent of the externality is required to match the transition path of output in South Korea. A series of quantitative experiments suggest that if the externality is removed from the model, then it cannot quantitatively match South Korea’s convergence pattern well. Chapter three documents a fact that that firms in developing economies face both financing constraints and face size-dependent distortions. The two distortions, however, affect firms in opposite ways. I build a model showing that the adverse effects associated with size-dependent distortions drastically reduce, and may even reverse, if firms also face financing constraints. This occurs because the misallocation effects of the two may offset each other. The quantitative analysis shows that size- dependent distortions estimated from data lead to up to 25 percent of output drop if they are implemented alone, but have virtually no effect on aggregate output in the presence of empirically relevant capital financing constraints.