MODELLING KENYA MACROECONOMIC INDICATORS USING PRINCIPAL COMPONENT ANALYSIS

Abstract

Kenya's economic growth has been lower than in other nations, especially the European nations for a long time. Macroeconomic indicators are the main factors that affect the economic growth of a country. The study sought to model Kenya macroeconomic indicators using principal component analysis. The study used PCA to capture data of 34 macroeconomic indicators for the period 1980 to 2019. The study aimed at applying PCA to reduce the dimensionality of the macroeconomic indicators and classify them into principal components. The study aimed at improving the way macroeconomic data has been handled in the past since several assumptions about the relationship between macroeconomic variables and economic growth have been made in the previous studies. The KMO statistics was found to be 0.720 and the p-value was 0.000. The KMO statistics indicated that the correlation matrix was appropriate for component analysis and the p-value depicted a significant difference. With reference to the correlation matrix, the variable were found to be closely correlated. Principal Component analysis was used to reduce the variables using varimax technique to principal components without compromising the variability of the original data. Only 8 variables (Principal Components) were retained since they explained 85% of the overall variations after scree plot, Kaiser Criterion and parallel analysis extraction approaches were utilized. The first component explained 28% of the total variance and was highly correlated with 10 macroeconomic indicators. Since the first principal component had the highest variance it was concluded that the monetary related macroeconomic indicators greatly impact economic growth in Kenya. Future researchers should consider having more diversified variables to help explain how economic growth is impacted by the all-round macroeconomic indicators