Janusz Brzeszczynski, Ph.D.1, Jerzy Gajdka, Ph.D.2, and Tomasz Schabek, Ph.D.1. (1) Department of Capital Market, University of Lodz, ul. Rewolucji 1905 r nr 41, Lodz, 90-214, Poland, (2) Department of Capital Market, University of Lodz (Poland), ul. Rewolucji 1905 r nr 41, £ódŸ, 90-214, Poland
This paper focuses on the intervalling effect in the estimation of beta parameter. We use the sample of all stocks listed on the Warsaw Stock Exchange in the period of 2003-2006 and analyze a wide range of intervals from 1 day to 21 days (i.e. one month). Whenever ARCH effects are detected, we employ ARCH models. The results clearly show that in the Polish stock market the interval effect does exist, i.e. the beta parameter for the same stock estimated using returns from various intervals is considerably different, and there is a tendency of beta estimates to decrease when the returns interval is shortened. The differences in the beta estimates are also apparent when ARCH effect is neglected and the models parameters are determined using a simple OLS methodology.