Garch modelling in share prices for specific companies in Kenya
Kinyanjui, Josphat k.
Mutiso, John M.
Omar, Latifa s.
MetadataShow full item record
The Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models of volatility that were introduced by Engle (1982) and Bollerslev (1986) are specifically designed to capture the volatility clustering of returns. This model is applied in modeling the heteroskedasticity in share prices for specific companies in Kenya. It is a key financial instrument for long term savings and for speculative purposes which is essential for both local and International investors. The volatility of financial asset returns changes overtime, with periods when volatility is very exceptionally high interspersed with periods when volatility is unusually low. From the results the mean and variance were more or less constant but it is not at all random (a strong seasonal pattern remains). The variation is seen in the years 2002/2003 and 2007/2008. The Kenyan economy undergoes a cycle of about five years which may either impact in a positive way or negatively on the share prices before stabilizing after duration of two years. The estimated GARCH models fit the data well, thereby confirming the empirical evidence in Bollerslev (1992), and that the GARCH (1, 1) is usually adequate in describing many financial time series.