PPP in the 34 OECD countries: evidence from quantile-based unit root tests with both smooth and sharp breaks

ABSTRACT

Conventional unit root tests have mostly failed to validate the PPP. Quantile-based unit root tests by previous research have provided some support for the PPP. In this article, we take an additional step and incorporate sharp shifts and smooth breaks into the quantile-based unit root test and re-examine the PPP in each of the 34 OECD countries over the period 1994:01–2016:03. We find support for the PPP in 18 countries of Austria, Chile, Estonia, Finland, France, Germany, Italy, Korea, Mexico, Netherlands, New Zealand, Poland, Portugal, Slovenia, Sweden, Switzerland, Turkey and the United Kingdom.

KEYWORDS:

• PPP
• OECD countries
• quantile unit root test
• sharp shifts
• smooth breaks

JEL CLASSIFICATION:

• F31
• F30

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ For a review article, see Bahmani-Oskooee and Hegerty (2009).

² Some recent examples are Taylor (2004), Taylor and Taylor (2004), Enders and Chumrusphonlert (2004), Bahmani-Oskooee, Kutan, and Zhou (2008), Kim and Perron (2009), Chang and Tzeng (2013), He and Chang (2013), Boero, Mavromatis, and Taylor (2015), Bahmani-Oskooee et al. (2015) and Baharumshah, Soon, and Wohar (2015).

³ For deviation of PPP, see Bahmani-Oskooee and Nasir (2005) and Sjölander (2007).

⁴ Note that Bahmani-Oskooee, Chang, and Wu (2014) modified the panel unit root test of Carrion-i-Silvestre, Del Barrio-Castro, and Lopez-Bazo (2005) by allowing for two different types of multiple structural breaks: first, breaks in the intercept (without linear trend), and second, breaks in the intercept and slope of the linear trend. We like to do the same for the time-series quantile unit root test.

⁵ Here is the web site for the data: http://www.bis.org/statistics/eer.htm.

⁶ Equation (2) is not only an extension of Enders and Holt (2012) but also a combination of Carrion-i-Silvestre et al. (2006) and Becker, Enders, and Lee (2006) tests.

⁷ Note that n and k are number of frequencies.

⁸ Enders and Lee (2012) also favour setting n = 1 in order to save the degrees of freedom and prevent the over-fitting problem.

⁹ Note that establishing validity of the PPP in some quantiles but not all amounts to testing the weak form of PPP. The concept is similar to establishing the validity of PPP within two structural break points or simply using observations from part of the sample (or part of the distribution in our case). For more details, see Belke, Beckmann, and Kühl (2011a, 2011b).

¹⁰ Rejection of PPP in the remaining countries imply that shocks to the real effective exchange rate has permanent impact on the rate such that once deviated, it does not converge back to its mean.