Detecting Multiple Breaks in Financial Market Volatility Dynamics

Table 8: Sequential sample segmentation approach of the Kokoszka and Leipus (2000) test for
detecting multiple breaks in stock market returns

SMI	Process Subsamples (observat.)	Kokoszka and Leipus Test
SMI	Process Subsamples (observat.)	Observ.	k Date	max. UT⁽^k ŋ Number σHAC of Breaks
FTSE	\|rt\| 1-3338	2197	5/6/97	5.861* 4
	1-2197	1021	2/12/92	3.508*
	2198-3338	2497	30/7/98	1.366*
	1021-2197	1631	5/4/95	2.041*
	( rt )² 1-3338	2239	4/8/97	5.265* 4
	1-2239	991	21/10/92	2.744*
	992-2239	1556	21/12/94	1.734*
	992-1556	1279	29/11/93	2.507*
HSI	\|rt\| 1-3338	2247	18/8/97	4.619* 6
	1-2247	1630	4/4/95	1.815*
	2248-3338	2554	19/10/98	2.818*
	1-1630	1239	4/10/93	2.416*
	1240-1630	1404	23/5/94	2.540*
	1405-1630	1534	21/11/94	1.953*
	( rt )² 1-3338	2249	18/8/97	2.181* 2
	2250-3338	2554	22/10/98	2.072*
NIKKEI \|rt\|	1-3338	2237	31/7/97	3.091* 5
	1-2237	966	16/9/92	2.476*
	2238-3338	2792	16/9/99	2.161*
	1-966	293	16/2/90	3.119*
	2238-2792	2365	27/1/98	1.716*
( rt )	² 1-3338	2295	21/10/97	1.972* 4
	1-2295	966	16/9/92	2.470*
	2296-3338	2861	22/12/99	1.757*
	1-966	295	20/2/90	2.162*
S&P500 \|rt\|	1-3338	2110	4/2/97	6.841* 4
	1-2110	776	30/12/91	4.143*
	2111-3338	2495	28/7/98	1.863*
	780-2110	1776	25/10/95	2.130*
( rt )	² 1-3338	2146	26/3/97	4.888* 3
	1-2146	779	31/12/91	3.159*
	780-2146	1814	18/12/95	2.343*

Notes: The test is based on a sequential sample segmentation approach with a 1% significance level for each segment.