Purchasing Power Parity in economies in transition: evidence from Central and East European countries

Abstract

The present study tests for the validity of long-run Purchasing Power Parity (PPP) for 17 European economies in transition. Analysis is performed following the methodological suggestions expressed in recent studies for PPP. Long-run PPP is initially tested for each economy vis-a-vis the USA, using the Johansen cointegration methodology and then for the whole set of countries using the Larsson et al. (2001) panel cointegration technique. The analysis provides support for long-run equilibria, but the coefficients of the estimated cointegrating vectors violate the symmetry and proportionality hypotheses suggested by PPP. We provide some arguments for these findings, based on the existing literature on transition and foreign exchange markets.

Acknowledgements

I would like to thank Nikos Mylonidis, the editor of the present paper and an anonymous referee for helpful comments and suggestions. All remaining errors are entirely my responsibility.

Notes

¹ In this, the categorization of Froot and Rogoff (1995) is followed.

² Dibooglu and Kutan (2001) identify the impact of nominal shocks on the nominal and real exchange rates in Hungary and Poland; Suppel (2003) illustrates the important effect of real shocks to the real exchange rates of Poland, the Czech Republic and the Slovak Republic; Borghijs and Kuijs (2004) show that it is mainly nominal shocks, which have contributed significantly to nominal exchange rate variability in five CEE Countries (the Czech Republic, Hungary, Poland, the Slovak Republic and Slovenia).

³ Most of these papers find evidence for relative PPP but fail to support strong PPP.

⁴ See Taylor (1988), for the derivation of Equation 3 in a model allowing for transportation costs and measurement errors.

⁵ The panel cointegration analysis uses a common sample length for all countries.

⁶ Their diagnostic tests do not indicate any serious mis-specification (serial correlation and/or non-normality) problem. They are not reported here for space reasons but are available on request.

⁷ They argue that in panel tests testing for the stationarity of real exchange rates, the non-stationarity hypothesis is rejected as a result of a few stationary real exchange rates in the panels. Still, alternative tests to overcome this problem turn out to be generally less powerful (see also Sarno and Taylor, 2002).

⁸ In any case, this becomes evident by the results referring to the individual system analysis.