ABSTRACT

 In recent times, it has become imperative to east a steady watch at one's nation's Currency or foreign exchange rate digit, as this rate is one of the major economic indicators. 1mm tue pFescnted study, observed data set being I he h'n'cign exchange rate of Naira, 200 -2006, is modeled using BoxJcnki us apiwoach to AR IMA iimodcling, and sonic recognized criteria are evaluated mu order to COII1J)IFC the diagmiostic clicekimig Oil lime likely models (as suggested by (lie au(oeorrclalion and partial autocoirelation plots ol original data set and the re';idtmals from the lifted trend) In arrive at one that gives the best lit to time data set. From the analysis in chapter Ihuec of this work, ii was discovered I hat the original data set Follow an S-cured fiar trend analysis, was not stat ional'y of which by diliercucing and/or (letrcndilig the original data, stationary was .mnet, ARIMA (1.0,0) without constant represented the data set vcll i residuals from the titled S-Curve are used for modeling AR IMA and flumimi ly, the residuals from Ike fitted models follow in approxi nualion the normality clist ,'ihut ion. It is FCCOI)) nuciidcd that the goVerummuelit should give more atleuution 011 the Naira. amid ellijJloy goo(i CCO1iOiiiiC and mnonclai'y policies, SO as to maintain its increase rather than decrease, however, achieve a realistic exchange mate of the Naira. In this study, the magnitude of conputational error is at the IllinilnhlIui as all time cowputatons were aided I)y time statistical packages: MINI 'IA Ii amid Mierosolt Fxccl.