Cointegration analysis of public debt in Kenya

Abstract

In this paper, we consider cointegration analysis in a VECM framework. More precisely, we analyse the macroeconomic indicators so as to identify the determinants of Kenya's public debt for the period 2001 to 2021 by applying the Johansen cointegration test coupled with VECM analysis. Quarterly time series data sourced from CBK are used. Public debt is the dependent variable while independent variables are USD exchange rate, capital and reserves, trade balance, budget deficit, net foreign assets, interest payments on debts and credit to the private sector. Firstly, the variables are tested for stationary using unit root test. Trade balance, budget deficit, and interest payments on debts are stationary in the levels while public debt, USD exchange rate, capital and reserve, net foreign asset, and the credit to the private sector are non-stationary, however, stationary at first difference making them integrating time series of order one at a 5\% level of significance. Secondly, the VAR model is estimated using OLS. The results of the ECM indicate cointegration relationship with error term of 0.0454. The ECM identifies net foreign asset, USD exchange rate, and capital and reserves as the main determinants of increasing public debt following a long-run relationship. Net foreign assets and credit to the private sector reduces public debt while USD exchange rate and capital and reserves increases public debt in a long-run relationship. Thirdly, the VECM model is statistically significant at a 5\% level. Finally, under the prevailing financial mechanism, public debt is projected to hit Ksh 9.453 trillion mark with a margin of error of 0.556 trillion by June 2023.