| Abstract [eng] |
In this paper we investigate the relationship among the biggest Baltic States construction sector companies daily returns. The sample period is from 2008-11-01 to 2016-05-06. An adequate VAR model is constructed after analyzing the stationarity of the variables and performing lag selection. It appears that one company's return values depend on foreign company's past values. Later GARCH and DCC models are used to study the conditional correlations among the paired firms. It is discovered that all companies returns correlate positively insignificantly, except Panevėžio Statybos Trestas and Latvijos Tilti returns, which correlate negatively. After construction of GARCH(1,1) and sGARCH(1,1) models for each of the company, it is found out that Panevėžio Statybos Trestas returns are the most sensitive to the external shocks. This sensitivity influences multivariate models: DCC models, where Panevėžio Statybos Trestas conditional correlation with other company's returns are analysed, have the highest \(\alpha\) parameter estimators (0.035074) and the correlations vary the most. |
