|
Article Excerpt
This paper utilizes the dynamic error-correction model (DECM) to examine the issue of purchasing power parity (PPP) for 11 developing countries (Argentina, Bolivia, Colombia, Cote d'Ivo ire, Ecuador, Guatemala, Kenya, Nigeria, Peru, South Africa, and Venezuela). For comparison purposes, evidence from the traditional unit root methods of the augmented Dickey-Fuller (ADF) and Phillips-Perron is presented. The results from the conventional unit root tests failed to find evidence of PPP in all of the cases. However, the results from the generalized error-correction model detected evidence of PPP for nine out of the 11 countries under consideration. Based on these results, it was concluded that PPP holds in the long-run for the sample countries and that the implicit restrictions associated with unit root tests prevented earlier studies from finding evidence in support of PPP theory. (JEL F31)
Introduction
The existence of purchasing power parity (PPP) has been a subject of considerable debate among monetary economists. The PPP theory (frequently referred to as the law of one price) stipulates that the exchange rate adjusts over time to accommodate inflation differentials between two countries. PPP is the cornerstone of the monetary models of exchange rate determination [Dornbucsh, 1976; Musa, 1982] and, as such, deserves the increased attention that it has received in the literature.
Most monetarists agree that deviations from the PPP frequently occur in the short-run. For example, Dornbusch [1980] and Frenkel [1978] found evidence against PPP in the short-run. However, the existence of PPP in the long-run, although widely researched, has produced mixed results in the extant literature. For instance, Abuaf and Jorian [1990], Darby [1983], Baillie and Selover [1987], Meese and Rogoff [1988], Mark [19901, Hakkio [19841, and McCloskey and Zecher [1984] found evidence of PPP in the long-run. In contrast, Cooper [1994], Messe and Singleton [1982], and Ahking [1997] found evidence against PPP in the long-run.
Meese and Singleton [1982] marked the turning point in the investigation of PPP. They found that nominal exchange rate has a unit root. This finding was interpreted as evidence that the nominal exchange rate follows a random walk, indicating that its impact is not meanreverting. In other words, changes in nominal exchange rate are expected to be permanent. Consequently, the long-run PPP hypothesis could not be confirmed. Since then, most studies have used the classical unit root methodology to test for PPP. In most of the cases, the existence of unit roots in the real exchange rate could not be rejected. For example, Adler and Lehmann [1983], using the conventional test, found that the real exchange rate follows a random walk and concluded that there is no tendency for the PPP to hold in the long-run.
Manzur [1990] and Manzur and Ariff [1995] investigated the existence of PPP for the Group Seven (G7) countries in conjunction with the ASEAN countries. They rejected the hypothesis of unit roots based on the results from the Sim's test. However, when they applied the Dickey-Fuller (DF), Augmented Dickey-Fuller (ADF), and Phillips-Perron (PP) tests, the hypothesis of unit root in the exchange rate could not be rejected. They therefore concluded that PPP holds in the long-run. Similarly, Abuaf and Jorian [1990] also found evidence in support of PPP in the long-run. Whitt [1992] investigated the hypothesis that the exchange rate follows a random walk using monthly data and Sim's test. He found that the real exchange rate does not follow a random walk, which indicates that PPP holds in the long-run.
Cooper [1994] investigated the validity of PPP by testing for stationarity and cointegration for Australian, New Zealand, and Singaporean currencies for the period 1973-92. He found that both the unit root and cointegration tests failed to detect evidence in support of PPP in the long-run for the three countries under consideration.
Ahking [1997] undertook a more sophisticated investigation of PPP by using the Bayesian unit root approach. Ahking found that there is little probability that the exchange rate and price level have a stable relationship in the long-run. Huang and Yang [1996] used both the Engle-Granger [1987] two-step approach and the Johansen [1988, 1991] maximum likelihood procedure to test the existence of PPP in the long-run. Interestingly, the Engle-Granger test rejected the hypothesis that PPP holds in the long-run. However, the Johansen test provided evidence in support of long-run PPP. When they re-estimated their data with the Monte Carlo simulations, they found that the Johansen procedure was biased in finding evidence of PPP because the normality assumption was violated most of the time.
Dorodian, Jung, and Boyd [1999] found that in the long-run, PPP tends to hold more often under a floating exchange rate regime than under a fixed exchange rate system. Lee [1999] used sample data from 13 Asia Pacific economies, which included developing and a few newly industrial countries, to test the validity of PPP in the long-run using the generalized error correction model. He concluded that PPP held in the long-run for the sample countries....
|
