This dissertation focuses on the financial markets including stock markets, commodity futures and options markets.

Chapter 2 studies the trading activity in commodity futures and options markets. Little is known about trading activity in commodity options market. We study the information content of commodity futures and options trading volume. Time-series tests indicate that futures contracts in a portfolio with the lowest options-to-futures volume ratio (O/F) outperform those in a portfolio with the highest ratio by 0.3% per week. Cross-sectional tests show that O/F has higher predictive power for futures returns than such traditional risk factors as the carry, momentum, and liquidity factors. O/F has longer predictive horizon for post-announcement returns than the information contained in the monthly World Agricultural Supply and Demand Estimates (WASDE) reports. The analysis of the weekly Commitments of Traders (COT) reports indicates that commercials (hedgers) provide liquidity to non-commercials (speculators) in short-term in commodity options market.

Chapter 3 explores what kinds of information can explain the USDA forecast errors in crop ending stocks. In the empirical analysis using Markov Chain Monte Carlo (MCMC) method, we find that the futures basis, level of monthly ending stocks, and level of planted area are significant to explain the forecast errors. The out of sample test is employed and the adjusted forecasts improve the forecast accuracy of crop ending stocks.

Chapter 4 investigates the liquidity effect in Chinese stock market using an asset pricing model. The empirical results show that liquidity has a significant effect on stock returns and the liquidity premium exists in Chinese stock market. However, neither CAPM nor Fama French three-factor model can explain the liquidity premium. We propose a new two-factor (market and liquidity) model in which the liquidity factor captures two dimensions of liquidity. The two-factor model performs well in explaining the liquidity premium. Furthermore, unlike CAPM and Fama-French three-factor model, the two-factor model is able to explain the size effect in Chinese stock market.