TWIN DEFICITS IN EAST AFRICA: A TIME SERIES ANALYSIS APPROACH

Abstract

Purpose of the study: The purpose of the study was to determine the validity of the twin deficits hypothesis in Kenya, Uganda and Tanzania.

Problem statement: In the past two decades, there has been a deterioration of balances in both the fiscal and current accounts in Kenya, Uganda and Tanzania. The twin deficits hypothesis proposes a positive causal relationship between fiscal deficits and trade account deficits. Many studies have researched on the existence of twin deficits. Some studies carried out have concluded that fiscal deficits improve current account deficits. It has therefore been difficult to obtain any relationship between the negative balances in the fiscal and current accounts. This may be because the studies did not allow for structural shifts and conditional heteroskedasticity. The study therefore sought to answer the question.

Study methodology: Data was analyzed using EVIEWS 8 software. This research adopted the use of time series data to understand the effect of negative fiscal balances on current accounts. The study incorporated structural breaks and conditional heteroscekedasticy into the time series data.  The study included a number of empirical tests based on various econometric techniques such as unit root tests; VAR-GARCH analysis. A two variable model of the VAR-GARCH model was used to establish the dynamics between the twin deficits. Lag length tests was also done to establish the appropriate lag lengths. The time series comprised of annual Government financial statistics of each country obtained from their respective treasuries.

Results of the study: the study revealed that that the UDMax methodology selects a multiple statistically significant break at the years; 1987, 1992, 1997, 2003 and 2010. The results revealed that Uganda data had more structural break points, among the three countries under study. Using the Kenya’s data, the goodness of fit (r squared) for the VAR-GARCH model was 49.54%. This means that fiscal balances explain 49.54% of the variation in current account balances. The relationship between the current account balances and fiscal balance (-2) is positive and significant (0.566497, p=0.0059).  Using the Uganda data, fiscal balances explained 74.04% of the variation in current account balances. The relationship between the current account balances and fiscal balance was insignificant. Tanzania’s data indicated that the goodness of fit (r squared) for the short run models was 74.04%. This means that fiscal balances explain 74.04% of the variation in current account balances. The relationship between the current account balances and fiscal balance (-1) is positive and significant (0.576578, p=0.0031).

Conclusion: the study concluded that the relationship between the current account balances and fiscal balance is positive and significant in Kenya and Tanzania while in Uganda there is no significant relationship between current account balances and fiscal balances. The governments been spending heavily on social programs aimed at improving the livelihoods of their citizens such as provision of free education to reduce the high illiteracy levels and free medical services.

Recommendations: The governments should establish policies aimed at regulating the amount of borrowings. These policies will ensure that the government does not exceed the set limits when it comes to borrowing both internally and externally. Such action will promote economic growth by reducing the amount of deficits. These deficits are a burden to future generations since they cause economic imbalances which in turn affect the economic development of every nation. The balances also make the countries to be heavily indebted

Keywords: Twin Deficits, Time Series Analysis & East Africa