Purchasing Power Parity Influence on Real Exchange Rate Behavior in Romania

Ghiba, Nicolae Article Purchasing Power Parity Influence on Real Exchange Rate Behavior in Romania CES Working Papers Provided in Cooperation with: Centre for European Studies, Alexandru Ioan Cuza University Suggested Citation: Ghiba, Nicolae (2011) : Purchasing Power Parity Influence on Real Exchange Rate Behavior in Romania, CES Working Papers, ISSN 2067-7693, Alexandru Ioan Cuza University of Iasi, Centre for European Studies, Iasi, Vol. 3, Iss. 4, pp. 524-536 This Version is available at: https://hdl.handle.net/10419/198144 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ C CE ES S W Wo or rk ki in ng g P Pa ap pe er rs s, , I II II I, , ( (4 4) ), , 2 20 01 11 1 524 PURCHASING POWER PARITY INFLUENCE ON REAL EXCHANGE RATE BEHAVIOR IN ROMANIA1 Nicolae Ghiba “Alexandru Ioan Cuza” University of Iaşi [email protected] Abstract: Purchasing Power Parity (PPP) represents a fundamental concept in exchange rate modeling. The main idea is given by equality between prices in two different countries when expressing in the same currency. This paper aims to analyze the behavior of real exchange rate between EURO and Romanian new leu (RON) under PPP paradigm. We use the Augmented Dickey-Fuller and Phillips-Perron stationarity tests in order to check real exchange deviations from PPP. Also, we investigate the existence of a connection between long-term between nominal exchange rate and industrial producer price indices from Romania and euro area. The main conclusions of this research highlight that PPP doesn’t holds; real exchange rate stationarity tests doest not confirm the stationarity, thus between the aforementioned three variables it doesn’t exists any equilibrium relation. Keywords: purchasing power parity, real exchange rate, stationarity, cointegration JEL Classification: F31, C32, E31 1. INTRODUCTION One of the most important theories in international finance refers to purchasing power parity. In absolute form, two price indices should have the same value after the conversion in the same currency. In its relative form, the theory reflects the equality between exchange rate modifications and price indices differential among countries. In reality, the price indices are elaborated using different products and weights, making the comparison difficult in being accurate. Also, the transportations costs, tariffs, taxes (or any other trade restrictions) and arbitrage operations influence the real exchange rate. The theory was elaborated by Gustav Cassel in 1932 and it was empirically developed until nowadays. The main studies focus on the real exchange rate stationarity tests (considered the “PPP strong form”) and on the cointegration of nominal exchange rate, a domestic price index and a foreign price index (considered the “PPP weak form). International finance theory reflects two puzzles about PPP validity: there is no consensus in obtaining similar conclusions about PPP in long term (the first puzzle) and the real exchange rate 1 ACKNOWLEDGEMENT: This work was supported by the the European Social Fund in Romania, under the responsibility of the Managing Authority for the Sectorial Operational Program for Human Resources Development 2007-2013 [grant POSDRU/107/1.5/S/78342]. C CE ES S W Wo or rk ki in ng g P Pa ap pe er rs s, , I II II I, , ( (4 4) ), , 2 20 01 11 1 525 has a higher volatility in short term, compared with a slower mean reversion adjustment in the longrun (the second puzzle). The empirical analysis on the Central and Eastern Europe emerging countries found arguments in sustaining PPP validity for Romania and/or the countries from region but also its rejection. In a recent study, Acaravci and Ozturk (2010) argue that real effective exchange rate is non-stationary if we neglect the structural breaks, otherwise the results being valid for Romania and Bulgaria. As methodology we used the Augmented Dickey-Fuller (ADF) stationarity test for the real exchange rate (the equation left-side) and the Engle-Granger methodology for the relationship between nominal exchange rate, domestic and foreign price index. The present paper is structured in six parts as following:  theoretical formulation and model development;  a literature review about the empirical evolution of the model and a brief Central and Eastern Europe countries presentation under PPP paradigm;  a presentation of the methodology that we used;  data sources and preliminary tests;  the most important empirical results;  concluding remarks. 2. THEORETICAL FORMULATION AND MODEL DEVELOPMENT The purchasing power parity has two forms:  an absolute form that reflects the same price for a good in domestic and foreign market when we are expressing it in the same currency: * ,, titti PNERP  , i=1,2,…,n, (1)  a relative form which implies an equality between exchange rate volatility and price indices: ti tti ti tti P NERP P NERP , * , 1, 1 *1,    , i=1,2,…,n, (2) where t NER is the nominal exchange rate at moment t and * ,tt PP represent the domestic price index, respectively foreign price index at moment t. C CE ES S W Wo or rk ki in ng g P Pa ap pe er rs s, , I II II I, , ( (4 4) ), , 2 20 01 11 1 526 The absolute form implies the relative form, but the influence isn‟t valid for the inverse relationship. The purchasing power parity is correctly elaborated if the internationally traded goods are perfect substitutable, there aren‟t transport cost and other impediments which can affect the international trade. PPP is important in analyzing the real exchange rate behavior. As we know, the real exchange rate (RER t ) can be computed as a multiplication between nominal exchange rate and the ratio between foreign and national prices: t t tt P P NERRER * * (3) In logarithmic form, the relationship can be written as: )log()log()log()log( *tttt PPNERRER  (4) If we use small cases, we get the next relationship: tttt ppnerrer  * (5) Empirical literature on the purchasing power parity relies on the relationship no. 5. PPP can be tested using real exchange rate stationarity test or cointegration methodologies: Engle-Granger (as a bivariate approach) or Johansen methodology (as a multivariate option) for the relationship between the nominal exchange rate, domestic price index and foreign price index). 3. PURCHASING POWER PARITY IN CENTRAL AND EASTERN EUROPE (CEE) EMERGING COUNTRIES. A LITERATURE REVIEW 3.1 Evolutions in empirical literature Early empirical PPP approaches (in ‟70s) had analyzed following relationship: tttt uppner  **  (6) Authors who studied the relationship 6 had applied coefficient restrictions tests: β=1, β*=-1. If the coefficients are equal and they have opposite signs we have a “symmetry condition”. If the coefficients are equal with the unity and they have opposite signs we get a “proportionality condition” (Frenkel, 1981). In the second part of „80s, economists developed the stationarity as a measure for the permanent deviation form purchasing power parity level. C CE ES S W Wo or rk ki in ng g P Pa ap pe er rs s, , I II II I, , ( (4 4) ), , 2 20 01 11 1 527 In the end of „80s, Engle and Granger (1987), followed by Johansen (1988) found a strong equilibrium relationship between variables using a new concept: the cointegration. The null hypothesis is given by the permanent deviations form linear combination between exchange rate and prices. Their methodologies were considering a success and it influenced empirical research until today. More appropriate to nowadays new approaches regarding purchasing power parity were been developed:  tests on longer time periods, which include various currency arrangements according to what happened in the international monetary system;  tests that are using panel data and stationarity procedures;  new econometric techniques based on nonlinearities. 3.2 Power purchasing parity evidences in transition countries One of the first PPP modern approaches of European Central-Eastern countries was made by Choudry in 1999. Analyzing a group of four countries (Romania, Poland, Russia and Slovenia) he found that PPP holds in its relative form for Slovenia and Russia. In an extensive analyze on a large group of countries from this region (Bulgaria, Czech Republic, Hungary, Poland, Romania and Slovak Republic in 1991-1998 period) Christev and Noorbakhsh, (2000) identified a long-run term relationship between prices and the exchange rate, despite the law of one price, proportionality and symmetry violation. Sideris (2006) examine the PPP validity for a long time horizon using cointegration for a 17 transition economies and find the same symmetry and proportionality violation, but also evidence of PPP validity in the long-run. Taylor and Sarno (2001) found an evident appreciation trend of the real exchange rate for the region‟s countries in „90s period. They consider that real shocks are more important than the nominal ones in exchange rate determination. Kim and Korhonen (2002) studied the PPP using real exchange rate stationarity for panel series in five countries: Czech Republic, Hungary, Poland, Slovak Republic and Slovenia, and argue that PPP doesn‟t holds. A similar approach for the transition countries finds that PPP holds (Solakoglu, 2006). Using the Johansen cointegration methodology in 1994-2000 periods, Barlow (2003) reached to a set of different results when combining the states between them. Koukouritakis (2009) analyzed PPP validity between twelve new member states of European Union (Bulgaria, Cypru, Czech C CE ES S W Wo or rk ki in ng g P Pa ap pe er rs s, , I II II I, , ( (4 4) ), , 2 20 01 11 1 528 Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Romania, Slovak Republic and Slovenia) by using Johansen cointegration test. He offers arguments in favor of a long-run equilibrium relationship between the nominal exchange rate and prices for Bulgaria, Cypru, Romania and Slovenia (Koukouritakis, 2009). Using non-linearity in PPP analyze, the theory is confirmed for Bulgaria, Croatia, Czech Republic, Hungary, Macedonia, Poland, Romania and Slovak Republic (Cuestas, 2009). A similar study on region find that PPP holds in Romania and Bulgaria, if we consider the structural breaks in the real exchange rate volatility (Acaracvi și Ozturk, 2010). 4. METHODOLOGY The purchasing power parity in a simplified form denotes that the modification degree of a currency is approximately equal to the difference between domestic and foreign price indices: tttt dpper  * , (7) where t er is the exchange rate in a logarithmic form and t p , respectively * t p are the logarithms of the national and foreign price index. t d denotes the deviations from purchasing power parity and it is associated with real exchange rate movements ( t rer ): tttt pperrer  * (8) Under these conditions, we admit that purchasing power parity holds in the long-run if the real exchange rate is a stationary series. A variable is stationary if it has a tendency in returning to a constant value. In other words, its trajectory must be around a mean value or around a linear trend. Economically, this means that any shock on series is temporary and it is absorbed in time. In practice, almost every variable is stationary and must be differenced. Hence, the exchange rate is nonstationary for the most cases and the series is first order integrated (requires just one differentiation). The econometric theory refers to a null hypothesis that claims a unit root in series. In our case, the real exchange rate is nonstationary. The most popular stationarity test were developed by Dickey and Fuller (ADF stationarity test), respectively by Phillips and Perron (1988). The difference between them is given by the less stringent restrictions on error process for Phillips- Perron test. These tests are important because it is necessarily for us to know the order of integration of our variables. If the obtained t-statistisc and associated probability reflect null hypothesis acceptance, than we conclude that purchasing power parity doesn‟t holds. C CE ES S W Wo or rk ki in ng g P Pa ap pe er rs s, , I II II I, , ( (4 4) ), , 2 20 01 11 1 529 Testing real exchange rate stationarity through Augmented Dickey-Fuller entails three assumptions: the intercept presence, the presence of an intercept and a time trend, and finally, the absence any deterministic element. For each supposition, we have build three different relationships: a) model A: includes both a drift and a linear time trend ttt tarerarer   ** 210 (9) b) model B: random walk with a drift ttt rerarer   10 * (10) c) model C: pure random walk ttt rerrer   1 * (11) For researchers, the most important coefficient is  . If its value equals zero, than the real exchange rate sequence contains a unit root (the series is nonstationary). The test estimates a regression equation using ordinary least squares, in order to determine an estimated value for  and associated standard error. In EViews 7, the associated probability indicates the rejection or acceptance of null hypothesis. Engle and Granger (1987) tested the cointegration between a set of integrated variables of first order: l(1). Their procedure is based on estimating the long-run relationship through regression, saving the residuals and testing their stationarity. If the the residuals are stationary, then the variables are cointegrated. Their stationarity shows that the purchasing power parity holds in the long run. In order to find relevant conclusions it is necessarily to use first order integrated series l(1) which need a single differentiation. According to Engle and Granger, if the * tt per  and t p are cointegrated, PPP holds in the long-run under following conditions (Enders, 2009):  between * tt per  and t p of the form * tt per  = tt p   * 10 exists a linear combination;  residuals ( t  ) are stationary;  variables have the same integration order. C CE ES S W Wo or rk ki in ng g P Pa ap pe er rs s, , I II II I, , ( (4 4) ), , 2 20 01 11 1 530 5. DATA SOURCES AND PRELIMINARY RESULTS After the socialist period, Romania “fought” with a highly inflation. The National Bank of Romania is trying to sustain a stable exchange rate in its goal of maintaining the prices stability in a domestic goods market characterized by many prices reported to the euro. During the recent crisis, inflation has grown with great effects on real exchange rate volatility compared with the nominal level. Among others, Romanian inflation rate were affected by the regulated prices – its level is appreciated at 21% in the consumer price index, but the regulation is generally applied to the nontradable part of the economy. According to Halpern and Wyplosz (2001), if over 10% of an economic sector is oriented to exports, we say that sector is tradable. The most prosper economic sector in this sense is the industrial one. PPP principles are based on the tradable part of the Romanian economy and, hence we consider that using an industrial producer price index is more adequate in our analysis. Purchasing power parity is analyzed for the period between 2000 (January) and the second half of 2011 (September) using monthly data taken from following sources:  National Bank of Romania Interactive Database for the nominal exchange rate between euro and Romanian new leu;  Eurostat for the industrial producer price index in Romania and Euro area. In our empirical analysis we use following abbreviations: Table 1- Abbreviations used in empirical analysis e_ippi Industrial producer price index in the Euro Area ro_ner Nominal exchange rate between EURO and Romanian new leu ro_ippi Industrial producer price index in Romania ro_rer_ippi Real exchange rate between EURO and Romanian new leu deflated with the industrial producer price indices d(*) Reflects the variable in first difference l*, log(*) Reflects the logarithmic form of the variable The evolutions of real and nominal exchange rate deflated with the industrial producer price index are reflected bellow, in the figure 1: C CE ES S W Wo or rk ki in ng g P Pa ap pe er rs s, , I II II I, , ( (4 4) ), , 2 20 01 11 1 531 85 90 95 100 105 110 115 120 00 01 02 03 04 05 06 07 08 09 10 11 e_ippi 20 40 60 80 100 120 140 160 180 00 01 02 03 04 05 06 07 08 09 10 11 ro_ippi Figure 1- Nominal and real exchange rate deflated with industrial producer price index in 2000M01- 2011M0 (EUR/RON) Source: graph made using National Bank of Romania Interactive Database for nominal EUR/RON exchange rate and Eurostat database for real exchange rate calculation Based on the graphic representation it seems that our variables (nominal and real exchange rate) follow a random walk process with no visible evidence of an explosive trajectory or a deterministic time trend. Before starting the empirical analysis we analyze the series characteristics in order to find the integration order. This information is useful in cointegration analysis, later developed in this paper. Figure 2 - Industrial producer price index in Euro area (e_ippi) and in Romania (ro_ippi) in 2000M01- 2011M09 period (2005=100) Source: EUROSTAT Using above graphical representation, both series look to be nonstationary. To be sure, we test this assumption using Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) stationarity tests. 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 4.4 4.8 5.2 99 00 01 02 03 04 05 06 07 08 09 10 11 ro_rer_ippi ro_ner