The Constant Status of Beta Essay

1639 Words 7 Pages
Beta being an efficient measure of riskiness of a security is an important financial instrument in investment decisions regarding estimation of market models, development of investment portfolios, estimation of cost of capital and emerging derivative markets. Since 1960s, the practical implication of CAPM has been in vague until recently challenged by some researchers Fama and French, (1992) for beta being insufficient in estimating the future returns of stocks based on historical data.

The constant status of beta in estimating stock returns is questionable by the academic researchers for the high-low variation in its parameters. As systematic risk is time variant in nature, thus it is necessary to consider beta as a time series process
…show more content…
From the time it has been developed, the Kalman filter technique has been the subject of extensive research and application in its original formulation (Kalman, 1960) where the state vector is estimated at discrete points in time.

1.1) Objective of the Study
The main objective of the study is to investigate the stochastic behavior of beta of 50 listed stocks on Karachi Stock Exchange Pakistan from January 1999 to December 2009 by employing state space model identified as modified Kalman Filter technique. This technique efficiently captures the unobserved variables in the data and estimates the variation in beta along with observed variables. If the Gaussian assumption is relaxed, the Kalman filter is deemed to be the best (minimum error variance) filter out of the class of linear unbiased filters as is evident from the study of Mergner and Bulla, 2005.

Bali, Cakici and Tang (2009) investigate the relationship of time-varying conditional betas based on the dynamic information available at any given point of time with firms and industry portfolios. The dataset comprises of daily returns of financial and non-financial firms listed on NYSE, AMEX and NASDAQ stock exchanges for the time period of July 1963 to December 2004. The authors have employed autoregressive, moving average and generalized autoregressive conditional

Related Documents