Current account sustainability for 21 African economies: Evidence based on nonlinear flexible Fourier stationarity and unit-root tests

We examine the mean reversion properties in the current account balance as a percentage of GDP under assumptions of smooth breaks and nonlinearity for twenty one African economies. Since there are reasons to indicate that the dynamic adjustment in the current account may follow a nonlinear process, we utilize a range of nonlinear stationarity and unit-root tests and compare across them to obtain a comprehensive picture of the pattern characterizing imbalances in part of the region. In particular, we apply the newly introduced Fourier stationarity test of Tsong, Lee, and Tsai (2019) in conjunction with Enders and Lee (2012a and 2012b) and Rodrigues and Taylor (2012) Fourier unit-root tests. However, rather than assuming a nonlinear current account adjustment, we test for “nonlinearity” using an approach that is robust to whether the current account series are stationary or integrated. We find strong evidence in favor of nonlinearity in eighteen current accounts out of the twenty one examined. Our empirical results show that the traditional linear and the widely used nonlinear unit-root tests of Kapetanios et al. (2003), Sollis (2009), and Kruse (2011) confirm sustainability of the current account balance for a small number of countries. Meanwhile, the Fourier based stationarity and unit-root tests confirm sustainability in a much larger group of countries.