期刊名称：Journal of Theoretical and Applied Computer Science
印刷版ISSN：2299-2634
电子版ISSN：2300-5653
出版年度：2013
卷号：7
期号：4
页码：56-69
出版社：Polska Akademia Nauk * Oddzial w Gdansku, Komisja Informatyki,Polish Academy of Sciences, Gdansk Branch, Computer Science Commission
摘要：This paper analyses the stock market linkages of the selected Central and Eastern European (CEE) markets (Czech Republic – PX, Hungary – BUX and Poland – WIG20) with the Western European stock market represented by the German DAX and studies also the co-movement between the individual CEE countries’ stock markets. The dynamic conditional correlation (DCC) models were used to model the co-movements and thereafter in some cases the smooth transition analysis was carried out in order to capture how these correla-tions evolve over time. The analysis was based on weekly data over the sample period Jan-uary 3rd, 1997 – November 29th, 2013 (883 observations). In the first step the asymmetric univariate autoregressive conditional heteroscedasticity model of Glosten, Jagannathan and Runkle (GJR) was estimated for individual stock return series. The results of the DCC-GJR models estimated in the next step show almost in all analysed cases the increasing lev-el of conditional correlations. In four cases (BUX_DAX, WIG20_DAX, BUX_PX and PX_WIG20) the DCC series were identified to be nonstationary – I(1) and nonlinear lo-gistic smooth transition regression (LSTR) model was used to capture the gradual transi-tion towards greater co-movements and to find out if the increasing level of DCC could be attributed to the accession of these countries into the European Union (EU) in May 2004.
关键词：stock market linkages; dynamic conditional correlation; logistic smooth transition regres-sion model; Central and Eastern European markets
