This dissertation is comprised of essays on econometric and asset pricing methods. In the first chapter, I develop a two-tier asset demand system that incorporates endogenous aggregate allocation and short sales, and propose a two-step estimation procedure with a novel instrument for aggregate estimation, which allows me to exploit both cross-sectional and time-series variation in institutional holdings. The estimated system provides a framework to answer questions related to demand-side effects of financial intermediaries and short sales in both aggregate and individual stock markets. I find institutional demand accounts for a large proportion, if not all, of observed return premiums in size, value and investment. The short leg, while increasingly important, cannot explain observed anomaly returns and the formation of the dot-com bubble. However, short sales do have significant yet disparate pricing impact on stocks with different characteristics. In the aggregate stock market, unobserved aggregate preference and beliefs rather than risk-return balance is the main driver of the return predictability of dividend-price ratio. In the second chapter, we develop an endogenous growth model with heterogeneous firms facing financial frictions, in which misallocation emerges explicitly as a crucial state variable. In equilibrium, misallocation endogenously generates long-run uncertainty about economic growth by distorting innovation decisions, leading to significant welfare losses and risk premia in capital markets. Macroeconomic shocks that affect misallocation are likely to have overly persistent effects on aggregate growth. Using an empirical misallocation measure motivated by the model, we find evidence showing that misallocation captures low-frequency variations in both aggregate growth and stock returns. Empirically, a two-factor model with market and misallocation factors prices size, book-to-market, momentum, and bond portfolios with an $R$-squared and a mean absolute pricing error close to the Fama-French three-factor model. In the third chapter, I propose a moment selection criterion for GEL estimation in moment-based models, facing possibly misspecified moments and/or many moments. The criterion, generalized focused information criterion (GFIC), trades off bias and variance in the limit as in finite samples, rather than allow validity to dominate. In the setting of a growing number of moments, GEL generally has preferable performance, as shown in the simulation study.