In models which allow for random jumps, statistical tests for jumps are typically non-standard and nuisance parameter-dependent. To handle these problems, we combine bounds and Monte-Carlo (MC) simulation techniques to derive nuisance-parameter-free bounds and obtain level-exact p-values for a wide class of processes with random jumps and time varying heteroskedasticity. When identified nuisance parameters are absent under the null, we show that MC p-values are finite sample, level-exact. To illustrate this easy-to-implement approach, we analyze the spot prices of four commodities (aluminium, copper, gold and lead) and the closing prices of four technology stocks (Intel, Microsoft, Oracle and Sun). We find significant jumps in these time series. Our approach can easily be extended to other nuisance-parameter dependent tests.

, , , ,
Journal of Economic Dynamics and Control

Khalaf, L, Saphores, J.-D. (Jean-Daniel), & Bilodeau, J.-F. (Jean-François). (2003). Simulation-based exact jump tests in models with conditional heteroskedasticity. Journal of Economic Dynamics and Control, 28(3), 531–553. doi:10.1016/S0165-1889(03)00034-4