Investigating Predictors of Inflation in Nigeria: BMA and WALS Techniques

Abstract

Abstract. The recent economic conundrum arising from the fall in the international oil price has threatened the maintenance of price stability, a key function of the central bank, therefore the need to investigate predictors of inflationary measures arises. The model averaging method considers uncertainty as part of the model selection, and include information from all candidate models. We analysed a wide spectrum of inflation predictors and all possible models for Nigeria CPI inflation using the Bayesian Model Averaging and Weighted Average Least Squares. The study uses fifty-nine (59) predictor variables cutting across all sectors of the Nigerian economy and three (3) measures of inflation, namely; all items consumer price index, core consumer price index and food consumer price index. (to be continued in page 302) Full Abstract (ENGLISH) The recent economic conundrum arising from the fall in the international oil price has threatened the maintenance of price stability, a key function of the central bank, therefore the need to investigate predictors of inflationary measures arises. The model averaging method considers uncertainty as part of the model selection, and include information from all candidate models. We analysed a wide spectrum of inflation predictors and all the possible models for Nigeria CPI inflation using the Bayesian Model Averaging and Weighted Average Least Squares. The study uses fifty-nine (59) predictor variables cutting across all sectors of the Nigerian economy and three (3) measures of inflation, namely; all items consumer price index, core consumer price index and food consumer price index. The results from both model averaging techniques showed that maximum lending rate, world food price index and Bureau de change exchange rate are the significant drivers of inflationary measures among focus variables, while foreign assets, credit to private sectors, net credit to government and real effective exchange rate are the drivers of inflationary measures, for the auxiliary variables, strongly supporting the monetarist and open economy views on inflation. The structuralist view is reported to be relatively weaker because government expenditure is only significant at 10.0 per cent.Résumé (FRENCH) L'énigmeéconomique récente découlant de la chute des cours internationaux du pétrole a menacé le maintien de la stabilité des prix, des principales fonctions de la Banque centrale, la nécessité d'enquêter sur les prédicteurs de mesures inflationnistes se pose donc. Le modèle avec une moyenne de méthode considère l'incertitude dans le cadre de la sélection du modèle et inclure les informations de tous les modèles de candidat. Nous avons analysé un largeéventail d'indicateurs de l'inflation et tous les modèles possibles pour le Nigéria IPC en utilisant le modèle bayésien en moyenne et pondérée moyenne des moindres carrés. L'étude utilise des variables prédictives de cinquante-neuf 59 coupe tous les secteurs de l'économie nigériane et 3 trois mesures de l'inflation, a savoir ; tout indice prix consommat...

Enhancing Nigerian Oil Price Forecasting: A Comprehensive Analysis of Model Averaging Techniques

“…It was found that certain versions of DMA prediction quality are higher than that of the naïve forecasting model. The methodology for the BMA [12][13][14]15,16,17,18,19] and DMA techniques investigate the important variables in a model selection process, and are presented in the next section of the paper.…”