Testing for Jumps in the Presence of Market Microstructure Noise

Piotr Pluciennik

Abstract

Discontinuous price changes called jumps are an essential component of financial asset price dynamics. As it was shown by Andersen, Bollerslev [1], Andersen et al. [2] and Lahaye et al. [3], jump occurrence in prices of various financial instruments is strongly correlated with macroeconomic announcements. Simulation researches show that the tests for jumps generally are sufficiently powerful, provided that high frequency data was used. Unfortunately, high frequency data is usually polluted by market microstructure noise (nonsynchronous trading, bid-ask bounce, discreteness etc.). In this paper we present results of test for jumps in the levels of three European stock indexes. We use two alternative approaches to testing of jump occurence by the assuption of presence of market microstructure noise: by Barndorff-Nielsen and Shephard [4] with, introduced by Andersen et al. [2], staggered bi- and tripower variation as estimators of integrated volatility and quarticity, and analytically modified form of swap variance tests introduced by Jiang and Oomen [5].