MULTIVARIATE ANALYSIS OF FOOD PRICES IN GHANA

ABSTRACT This study is concerned with the evaluation of the price levels of food items in various markets across Ghana. The objective was to determine the main dimensions along which to identify specific markets in which items are high-priced and those that are low-priced. To this end, data on prices in 2008 of fifteen food items in 100 selected markets from all the regions of the country were obtained from the Statistical, Research and Information Directorate of the Ministry of Food and Agriculture. The data thus obtained were multivariate in structure since prices of 15 food items were recorded from each of 100 markets. Principal Component Analysis, which is a procedure for data reduction and summarization, is considered appropriate for analyzing this high dimensional data set. Initial exploration of the data, using zero-order and partial correlation analysis and eigen-analysis among others further informed the choice of the technique. It was found that there are two important dimensions along which the levels of prices can be determined. These are the weighted sum of all the food items and the weighted sum of only food items that are considered as the main constituent of a typical local diet. A cluster analysis of the first component scores revealed that only a single market, Tepa in the Ashanti Region, was generally the lowest-priced market in the country. However, markets that are predominantly located in the Northern, Upper East and Upper West Regions are generally high-priced specifically the Bongo market in the Upper East region. There are, however, not much extreme-priced markets on the second component. The study shows that it will be economically beneficial to be mindful of the location and the size of the market from which one intends to make purchases of certain categories of food items.