Revisiting purchasing power parity in African countries: panel stationary test with sharp and smooth breaks

Abstract

In this study, we apply the Panel Stationary test with both sharp and smooth breaks to test the validity of long-run purchasing power parity (PPP) for 20 African countries using quarterly data over the period 1971I–2012IV. Empirical results of Panel stationary tests with both sharp and smooth breaks indicate that PPP holds true for 10 out of these 20 African countries (i.e., Burkina Faso, Cameroon, Ghana, Kenya, Niger, Senegal, Seychelles, South Africa, Tanzania and Togo) studied.

Keywords:

• PPP
• Africa
• panel stationary test
• sharp and smooth breaks
• Fourier function

JEL Classification:

• C21
• F31

Notes

¹ For some other univariate tests see Bahmani-Oskooee et al. (2008).

² With some exceptions, we found that the unit root null hypothesis was rejected for Cote D’Ivoire, Niger and Togo when both the ADF and PP tests were conducted. The KPSS test also fails to reject the stationary null hypothesis for Cote D’Ivoire and Niger (see Table 1).

³ Quarterly real effective exchange rates for each country come from Bahmani-Oskooee and Kones (2014).

⁴ Maximum numbers of breaks and frequencies have been fixed at 7. We report the results for break dates in a table which is available upon request.

⁵ For some other PPP and exchange rate related studies, see Taylor (1988), Hojman (1989), Bleaney (1992), Beach et al. (1993), Moosa (1994), Jung (1995), Huang and Yang (1996), Ahking (1997), Nachane (1997), Apergis (1998), Taylor et al. (2001), Holmes (2002), MacDonald et al. (2002), Arize et al. (2003), Peel and Venetis (2003), Baffoe-Bonnie (2004), Carrion-I-Silvestre et al. (2004), Enders and Chumrusphonlert (2004), Horne (2004), Jenkins (2004), Narayan (2005) and Sjölander (2007).