Detecting Structural Breaks in Some Major Macroeconomic Variables in Ghana

Abstract

This research was conducted to detect structural breaks in some major macroeconomic variables using single and multiple breaks approaches. The data used was obtained from the Central Bank of Ghana (BoG) and included monthly Consumer Price Index (CPI), 91-day Treasury bill (91 T-bill), Inflation rate, Cocoa Price (Cocoa P) and Crude Oil Price (Crude oil P) spanning from January, 2007 to December, 2012. The study showed that all the macroeconomic variables were not normally distributed as they were platykurtic in nature with their standard deviations far from their means, indicating volatility. Using ADF test, all the variables were not stationary at level, but stationary after first differencing. The Chow, Quandt Likelihood, Cumulative Sum test (CUSUM), Cumulative Sum Squared test (CUSUMSQ) and Bai and Perron multiple tests were used to check for structural breaks. They all showed structural breaks in all the variables. The observed variables were categorized into subgroups with smaller sizes of 72, 24, 12 and 6 months. The number of breakpoints detected by the Bai and Perron multiple tests were six (6) in all the macroeconomic variables and in each subgroup too. Detected breaks date from the CUSUM test were substituted into the Chow test and there were inconsistencies.

Keywords: structural breaks, macroeconomic variables, stationary