This thesis refines the measurement and prediction of liquidity in bond markets. In the first part, we adapt commonly employed liquidity measures to account for the unique nature of over-the-counter bond trading. We use these adapted measures to reevalute the impact of liquidity on bond prices. In the second part, we develop an approach to predict individual bond liquidity and analyze effects of expected bond liquidity. Finally, the third part introduces statistical tests to compare forecast model performances in the context of large panel data.