| Chapter 1 introduces a new nonparametric jump test for continuous-time asset pricing models. It distinguishes actual arrivals and different sizes of jumps. Asymptotic distribution of the test statistic is provided, and we demonstrate that it is desirable to use high-frequency data. We also discover false jumps in multiple jump tests, which have been ignored in jump intensity estimation. They create bias in the intensity estimates and lead to systematic loss in portfolio management and option mispricing. We offer a remedy for the error. Empirical evidence of false jumps in foreign currency exchange markets is also presented.; In Chapter 2 we study the impact of market microstructure noise on distinguishing jumps in continuous-time asset pricing models with the nonparametric jump test introduced in Chapter 1. A modified version of the test that accounts for noise is suggested. We show its asymptotic null distribution is the same as if noise is absent. Our simulation study supports our contention for the noise with finite variance, serial correlation, and cross-correlation with efficient prices. We explore dynamic jump intensity structure through the test, and find empirical evidence of jump clustering in foreign currency exchange markets. Its implication for financial risk management, in particular value at risk, is also discussed. |
