Detecting jumps amidst prevalent zero returns: Evidence from the U.S. Treasury securities

We examine the performance of conventional jump-detection methods for the U.S. Treasury notes. We first document how the Treasury market is different from the stock market: each day the Treasury notes have a large proportion of zero returns, because the vast majority of trades are executed at the best ask/bid quotes and spreads are mostly set close to the minimum tick. Moreover, the proportions of zero returns rather capture liquidity in the Treasury market. Given the distinctive feature (frequent zero returns) in the U.S. Treasury market, we find that conventional jump-detection methods are vulnerable to biases, leading to falsely identifying jumps. We propose a low-cost solution to the biases, and empirically support the arguments by using the actual data on the Treasury notes and macro-economic news announcements.