| dc.description.abstract | Value-at-Risk is an important concept in financial management, financial reporting and risk management. In this study, we
have used this tool to assess risk in stocks listed in the Nairobi Stock Exchange. It is commonly used because it summarizes risk into a
single value which is easily understood. Daily average share prices from January 2003 to December 2013 of Kakuzi and BAT stocks
were analyzed. This is a sample of the NSE-20 index stocks which are among Kenya’s top stock over 2008-2012 period. In estimating
VaR, we need to estimate the volatility of the returns, specify the holding period and the confidence interval.. We modeled the volatility
of the two selected stocks using a GARCH model. We selected the appropriate order of GARCH for each of the stocks using AIC instead
of using the most commonly used GARCH (1, 1). GARCH (4, 2) best fitted Kakuzi data and GARCH (5, 4) BAT data. From the residual
analysis the models performs well and we therefore used them in estimating VaR of each of the stocks. Backtesting a VaR model is
important as it helps determine whether the model is able to capture risk well. This study reveals after backtesting the VaR model of the
two stocks, that the model does not capture risk well since the actual number of exceedances exceeds the number of exceedances
proposed by 95% confidence interval.
Keywords: Risk, GARCH, VaR, Volatility, Stock | en_US |
