Journal de l'économie mondiale

Structural Breaks in Consumption of Tobacco in Kenya

Abstract

Edwin Kipyego Kipchoge*, Betty Korir, Argwings Otieno

The linear relationship between time series variables is normally assumed to be stationary or simply the mean and variances do not change over time and the covariance of a variable and its lagged values are constant over time. In real situations there are some major events that can influence the fluctuations of such time series variables. Structural changes can identify whether the linear relationship between independent and dependent variable changes over some time t. The main purpose of this paper is to highlight some of the major structural breaks in tobacco consumption in Kenya for the period 1980 to 2016. The data used was extracted from Kenya National Bureau of Statistics (KNBS) and Government statistical abstracts. The independent variables are the prices of tobacco both before and after tax and per capita income. The Zivot-Andrews test was applied to test one structural break, but Clement-Montane and Reyes test was used to identify the multiple structural changes. The results obtained were presented in a table form and figure for visual identification. Zivot Andrews showed that there were major and significant structural breaks in tobacco consumption in the year 2007(t-statistic value is -6.366), prices before tax in the year 1999(t-statistic value is -4.882) and per capita income had a break year 2006 (t-statistic value is -2.627). Clement Montane and Reyes results indicated that the variables under study had significant break years (p-values were less than 5 per cent significance level). Based on the results obtained, the consumption of tobacco had two major break structural breaks in the year, 1994 and 2003, similarly the income per capita break years were 1999 and 2009, also, the prices before and after-tax break years were in 1996 and 2003, and in 1995 and 2009 respectively. This structural change is an important observation in analysis as it improves the efficiency of the estimates and assist researchers and governments to identify policies that can help reduce tobacco use, reduce diseases associated with tobacco for a functional healthy nation.