The study investigates evaluate properties and performance of long memory models from emerging foreign markets return innovations between 1991 - 2020. The purpose of the study includes; investigate the persistence of shocks in Nigerian international markets, model long-range dependence, test the efficient markets hypothesis using fractionally integrated volatility models, develop an appropriate long memory model for Nigerian international markets, compare the advantages between short and long memory models in modeling for the returns in Nigerian international Markets and Give forecast values for future occurrences. The design for the study was an ex post facto research design. The data used for this study were Nigerian crude oil prices (Dollar per Barrel), exchange rate, and Agricultural Commodity prices extracted from the website of the Central Bank of Nigeria (CBN) www.cbn.ng. The total data points were 1044 and it spanned from 1st January 1991 to 30th January 2020. The statistical software used for data analysis was STATA 15 and OX metrics version 7. In an attempt to achieve the aim of the study, parametric and non-parametric methods of detecting Long Memory were applied. The study applied short and long memory models in an attempt to spot out the deficiencies associated with the short memory models. The results confirmed the presence of long memory in sales and returns on prices in Nigerian international markets. The presence of long memory in both sales and returns on prices in Nigerian International markets disprove the efficient market hypothesis which says that the future returns and volatility values are unpredictable. Similarly, base on performance evaluation using the Akaike information criteria, ARFIMA(1,-0.021,1) model was found to be the best fit model to the data after checking the adequacy of the model selected. Sequel to the above, it was recommended that there is a need for a strong financial and economic reform policy to curb persistent shocks in Nigerian international markets. This is because a stable local financial currency builds confidence in an economy, especially when foreign investors intend to invest in the country’s economy. For example, exchange rate policies also trim down the desire for local investors to trade in the international market. Also, for empirical estimation of long memory sales and returns on prices in Nigeria international markets, ARFIMA(1,-0.021,1) model should be considered appropriate. Two years (January, 20 to Dec-22) step ahead forecast shows that the predicted value for Cocoa Bean Sale using the ARFIMA (1,-0.021,1) falls between the range of 1.907247 to 1.915947.
Evaluating Properties and Performance of Long Memory Models from an Emerging Foreign Markets Return Innovations
