engleski

Testing Exchange Rate Determinants in the Framework of Vector Error Correction Modelling: Overvaluation and Tourism

The empirical evidence for either PPP or UIP individually is mixed at best. This paper combines the two arbitrage conditions into a single relationship, as the empirical literature is more supportive of such a combined relationship than of either PPP or UIP separately. Johansen’ s cointegration method is used to verify the existence of a long-run relationship between the HRK/EURO exchange rate, Croatia and EU prices, and Croatia and EU interest rates. Using the estimated relationship, we are able to determine whether prices, interest rates and the exchange rate were consistent with PPP and UIP over the sample period. We then calculate the equilibrium exchange rate that would have restored PPP and UIP – this is the equilibrium exchange rate HRK /EURO according to PPP and UIP. At the end of the paper we are able to calculate the elasticity of tourism demand in Croatia on each concept of nested equilibrium exchange rate. Since the PPP is a long-run condition, it is assumed that PPP forms the basis of expectations in the UIP condition. Algebraically, this relationship is obtained yielding: = Rearranging: (5) where Equation (5) can be assumed to represent the equilibrium condition. In the real world, however, nominal exchange rates are not, always and everywhere, determined by price levels and interest rates. For example, speculative activity or commodity price movements could lead to a sustained and significant deviation from equation (5). Therefore, macro-economists and policy makers are keen to know rather equation (5) can be considered as an equilibrium condition toward which exchange rates, price levels, and interest rates tend to move in the long run. In other words, whether price levels, interest rates, and the exchange rate are cointegrated or nominal exchange rates could be expected to deviate from this equilibrium condition, such that: (6) Where t has zero mean and finite variance and represents the deviation from equilibrium PPP-UIP condition. In the next section, the study empirically estimated equation (6), using multivariate cointegration approach to test for long-run equilibrium relationship. We estimate a cointegrating relationship between the nominal HRK/EURO exchange rate, Croatia and EU prices, and Croatia and EU interest rates. This involves estimating a Vector Error Correction (VEC) model. Formally, this study, using the Johansen technique, tests whether there exist one or more vectors of coefficients such that: w1e12t + w2p1t + w3p2t + w4i1t + w5i2t + w6 ~ I(0) 1 1 -1 - The expected values of w1, … ., w6, implied by equation (6) are displayed underneath. We start with a five-variable system Yt = [e12t, p1t, p1t, i1t, i2t]. In the system, p1 and p2 are the logs of the Croatia and EU consumer price indices, respectively, e12 is the log of the HRK/EURO exchange rate, i1 = log (1+ R1/100), i2 = log(1 + R2/100), R is the 3 month time deposit rate in Croatia(CROL) and R2 is the 3 month LIBOR rate (EURO). All series are seasonally unadjusted. The sample period is from 1994:1 to 2007:3 covering the recent floating exchange rate regime and domestic and international capital market liberalization. We try to descript the main results in non-technical voice because the complex technical analysis of cointegration is performed in very rigorous sense in the paper. First, this paper has provided evidence of a systematic long-run relationship between the exchange rate, interest rates and prices. The obtained result indicates that Croatian nominal exchange rate is, very closely in line with its explanatory variables. This relationship does match the strictest form of PPP and UIP, but even more stricter form of UIP and weak form of PPP since Croatia interplay of volatability among domestic and Euroland interest rates appear to affect the exchange rate more than relative prices affect the exchange rate. So, the main result of our study is that the estimated relationship used strict UIP and weak PPP, can estimate a conditional equilibrium nominal exchange rate by minimal standard deviation (or by the best performs). Second, the analysis has provided evidence that there is a cointegrating relationship between the exchange rate, interest rates and prices implying that the exchange rate is determined by combined two- fundamental arbitrage conditions viz., UIP and PPP. This suggests that prices (domestic and foreign) and interest rates (domestic and foreign) are playing significant role to derive the current exchange rate. Fluctuations in currency value HRK/EURO can be prevented, we would underline, first, by controlling interest rate movements and second, something less important, but implicitly, through achieving price stability as well. Third, the first cointegrating relation is normalized by the nominal exchange rates for Croatia and we found that the estimates of the normalized cointegrating vector have signs that match the theory of combined PPP and UIP, although the benchmark unrestricted conditional equilibrium exchange rate produced maximal misalignment and therefore maximal deviation in the long run from the nominal exchange rate. But nevertheless equilibrium exchange rate in other two combined forms follow very closely the trajectory path of nominal exchange rate – we didn't notice any relevant overvaluation accordingly restricted cointegration relations. We find that the elasticity of tourist demand on exchange rate fluctuations is calculated by theoretically accepted sign (beside when regressed on unrestricted conditional exchange version – even insignificant) and is very elastic. Overall, the findings have important implications for the enlivening the tourism demand because financial reforms and interest rate policy could provide help in designing effective exchange rate policy, and strengthening tourism position in Croatia. Pfaff, Bernhard (2006), Analysis of Integrated and Co-integrated Time Series with R, New York NY: Springer Johansen, S and K Juselius (1992), "Testing structural hypotheses in a multivariate cointegration analysis of the PPP and the UIP for UK, " Journal of Econometrics, 53, pp 211 – 44. Aysun, G. and Özmen, E. (2001), "Do PPP and UIP Need Each Other in a Financially Open Economy? The Turkish Evidence", Working Paper No. 101, Economic Research Center, Middle East Technical University. MacDonald, R (2002), "Modelling the long-run effective exchange rate of the New Zealand dollar", Reserve Bank of New Zealand Discussion Paper DP2002/02.

Purchasing Power Parity; Uncovered Interest rate; Vector Error Correction Modelling; tourism demand

Sažetak rada objavljen je u Knjizi produženih sažetaka, a puni tekst rada CD-u, Track No. 14, Economics of Business in Transition Economics