This thesis consists of three essays. The first two essays present empirical studies in which option market features related to information flows are examined. The third essay introduces a theoretical model to explain predictable dynamics in option pricing through the agents' learning process. In the first essay, I investigate the previously unexplored effects of asymmetric information on the adoption process of new equity options introduced into the market. I use a microstructure model to estimate measures of informational asymmetries. I discover that high informational asymmetries in the year prior to option listings produce larger levels of option adoption. Additionally, I find that option introductions induce reductions in asymmetric information. I also report that option bid-ask spreads start from low initial levels with a tendency to increase over time, which is unexpected since the introduced options are initially illiquid; however, this can be explained by the low level of initial activity by informed agents.The second essay examines whether the dynamics of the implied volatility surface of equity options contain exploitable predictability patterns. The option pricing predictability is expected due to the learning behaviour of agents in option markets. In particular, I explore the possibility that the dynamics of the implied volatility surface of individual equity options may be associated with subsequent movements in the volatility surface implicit in S&P 500 index options. I present evidence of strong relationships in the cross-section and the dynamics between implied volatility surfaces of equity options and S&P 500 index options. Moreover, I show that the predictability patterns of equity options are better characterized by the incorporation of information from the recent dynamics in the implied volatility surface of S&P 500 index options. Additionally, I analyse the economic value of the equity option predictability through trading strategies using straddle and delta-hedged portfolios, which produce abnormal risk-adjusted returns.Finally in the third essay, I introduce an equilibrium model to explain predictability patterns in option pricing through the learning process followed by investors. In this model the unknown fundamental dividend growth rate is subject to breaks, where the time periods between breaks follow a memoryless stochastic process. Immediately after a break there is insufficient information to price option contracts accurately. Therefore, a representative Bayesian agent has to learn step by step as new information arrives regarding the new fundamental value. I show that learning makes beliefs time-varying, which produces dynamic biases in option prices and implied volatilities. In addition, I find that learning generates different dynamic impacts on option contracts across moneyness and time-to-maturity; and hence it induces dynamics on the implied volatility surface. Furthermore, similarly to the predictability features observable in option market data, learning mechanisms make the option pricing dynamics predictable.