This Thesis is devoted to better understand market dynamics and asset pricing anomalies. In Chapter 1, which is co-authored with Andrea Hamaui, we study the effect of investors’ market expectations on asset pricing. Given traditional stock returns factor modelling and the prominence of the market factor, beliefs about market re- turns represent a natural primitive for expectations of stock prices. As the desire to increase market exposure generates excess demand for high beta assets from con- strained investors, we connect mutual funds’ expectations to the beta (or low vol) anomaly. We show that the beta anomaly is particularly strong for stocks purchased by over-optimistic mutual funds. On the empirical side, we first introduce a mutual fund-level measure of market expectations and confirm the model’s predictions for asset prices. In Chapter 2, which is co-authored with Andrea Hamaui, we study mutual funds’ trading behavior. In particular, we introduce the concept of "core" vs "satellite" holdings and we characterize positions depending on their longevity and interim re- turn in a fund’s portfolio. We show that core positions are relatively protected from selling in times of distress, as managers consolidate their portfolio. Next, we show that this theory has implications for asset prices and liquidity: core positions incur less downward contemporaneous price pressure as a result of outflows and are rela- tively more liquid. A behavioral model rationalizes those findings and validates the use of interim return and longevity as proxies for the "coreness" of a position. In Chapter 3, I develop a three-period asset pricing model with heterogeneity in firms’ size and a government that introduces a policy distortion. I find that large firms can better hedge the political uncertainty associated with this policy change through lobbying, which leads them to earn lower expected returns. I provide two strands of empirical evidence consistent with the model predictions. The first one looks at the behavior of a blue versus red industries around the unexpected results of the 2016 US Presidential election. The second one forms a political risk factor using a matching procedure, and shows that lobbying is indeed associated with a lower exposure to this factor.Ph.D.