| dc.description.abstract | Uncertainty in agricultural food markets generate such price volatility that, in
calculating measures of market-wide inflation for guiding monetary policy, a “core”
measure of inflation that removes food prices is preferred. This prompts the search for
knowledge, awareness and deep understanding of the trends, stationarity, cointegration and causality in the agricultural food prices in the markets, what this study
achieved objectively.
Twenty five food commodities were grouped into three categories; carbohydrates,
proteins and vitamins and minerals. The average monthly price per kilogram was
computed for each category, yielding sixty data points for each category. The data
was analyzed using the R statistical software to investigate their cointegration and
causality.
Based on Akaike Information Criterion, the optimal lag for the price time series was
found to be one. Based on this lag and using the Augmented Dickey Fuller test, the
Carbohydrate prices were found to be non-stationary; therefore they were differenced
then de-trended so as to achieve stationarity. The protein and Vitamins and minerals
prices were found to be stationary.
Using the Johansen test of cointegration, it was found that there exist two
cointegrated models of the food prices. Further, the Granger tests of causality
indicated significant unidirectional causality among the food prices. The prices of
proteins cause those of carbohydrates and those vitamin and mineral foods. The
prices of vitamin and minerals cause those of carbohydrates but not those of protein
foods. The prices of carbohydrate foods do not cause those of either proteins or
vitamin and mineral foods. Therefore, the recommended model is that which uses the
prices of protein and vitamin and mineral foods to explain the prices of the
carbohydrate foods. That is, Carb = f (Prot, Vit.Min). | en_US |
