Author(s): Tadele Kassahun Wudu and Salie Ayalew
Background: Floriculture has become one of the important high value agriculture industries in many countries of the world. The floriculture sector is a relatively new established sector in Ethiopia.
Methods: The aim of this study was to model the export sales of Tana flora for the period of January 2012- December 2017. The paper employs different univariate specifications of the Generalized Autoregressive Conditional Heteroscedastic (GARCH) model, including both symmetric and asymmetric models. The prediction of a future value used forecasting performance accuracy.
Result: The empirical results show that the conditional variance (volatility) is an explosive process for the price of flower export series, First differencing was used for transforming non stationary time series to stationary ones. Seasonal index was used to remove seasonal effects of the data. Additionally, the asymmetric GARCH models find a significant evidence for asymmetry in monthly export sales of flower series. The asymmetric Garch model, the EGARCH (1, 2) model under GED distribution was an appropriate model for the export sales of flower series. Furthermore, the out-sample forecasts indicated that the export sales of flower of the company has maturated and the sales function started to decline.
Conclusion: The comparison was made between the symmetric and asymmetric GARCH models, the relative mean squared error and mean absolute error measures, the empirical result suggests that EGARCH, EGARCH (1, 2) model under GED was good enough to describe and predicted the export sales of Tana Flora industry models are superior for EGARCH volatility forecasting.