Exchange Rate Volatility and Foreign Trade in CEEC

Exchange Rate Volatility and Foreign Trade in CEEC Ing. Lucie Tomanova, Finance, School of Business Administration in Karvina, Silesian University in ...
Author: Horatio Carroll
2 downloads 1 Views 393KB Size
Exchange Rate Volatility and Foreign Trade in CEEC Ing. Lucie Tomanova, Finance, School of Business Administration in Karvina, Silesian University in Opava, [email protected] Abstract The exchange rate plays an important role in a country’s export performance and currency volatility has impact on international trade, the balance of payments and economic performance, however, views on the impact of exchange rate volatility on international trade flows are inconsistent, thus it is necessary to examine this matter further, and with knowledge of the application to small open economies. This paper analyzes impact of exchange rate volatility on the export performance of Central and Eastern European countries. Using monthly time series data, the empirical analyses has been carried out for the period 01/1999 to 03/2013. Volatility’s impact on export performance is estimated on bilateral export flows of Czech Republic, Slovakia, Hungary and Poland to euro area. For the volatility measurement, G/ARCH models are used. Autoregressive distributed lag and error-correction approach are used to examine the impact of exchange rate volatility on the exports. The results suggest no significant relationship among the exchange rate volatility and export performance in CEE countries, impact of exchange rate volatility turns out to be ambiguous.

Key Words exchange rates, export, volatility, CEEC, VECM, ARDL

Introduction With inflation targeting and freely floating, Central and Eastern European countries (CEEC) exporters have faced during past years rather volatile exchange rates which may have influenced their behaviour. Countries which are subject of research are Czech Republic, Poland and Hungary, altogether with Slovakia forming Visegrad Four group and members of EU since 2004 with commitment to adopt euro as a single currency. This commitment was fulfilled so far only by Slovakia where euro was adopted in 2009. Regarding remaining countries, entry to the euro area will bring benefits, such as reduction in transaction costs, access to the large euro capital markets and also elimination of exchange rate risk. The relationship between exchange rate volatility and international trade and perception that higher exchange rate risk reduces international trade also helped motivate for single currency adoption. However, being a member of large and diverse euro area also poses challenges notably that monetary conditions will not always conform to the national requirements (OECD, 2004). In this regard, CEE countries are aware that nominal exchange rate adjustments are no longer in use and monetary policy will be delegated to European Central Bank, whilst economies will have to adjust without possibility of using monetary policy in their favour. Exchange rate volatility became important in the context of integration to euro

area as well. This paper investigates to what extent high exchange rate variability can be responsible for the developments in CEEC export volumes, i.e. what is the impact of exchange rate volatility on CEE countries export to euro area. In addressing this question, G/ARCH models were used as a measure of exchange rate volatility. As shown in the report of WTO (2011), in the last decade, research has moved from volatility rather toward the real exchange rate level and aggregated data from the data on the bilateral level. In this respect, the exchange rate depreciation stimulates exports and reduces imports, while appreciation is detrimental to exports and imports improve (Abeysinghe and Yeak, 1998). However, it is necessary to distinguish between short and long periods. Economic theory suggests that longterm trade flows are not affected by exchange rates, since relative prices do not change. Longterm effects are largely confirmed in models that deal with market failures, information asymmetries and product market failures (Haddad and Pancaro, 2010). In the short term movements in nominal exchange rates may change relative prices and thus influence international trade flows. These short-term effects on foreign trade are again not clear, because they are conditioned by the specifics of the economy, including the currency in which domestic producers invoice their products and structure of trade. The effect of depreciation can be completely different if domestic producers of goods valued in buyer’s currency, or in one of the world's currencies. This situation leads to lower real effect because it does not stimulate exports, but leads to import reduction (Staiger and Sykes, 2010). The views of exchange rate volatility effect on international trade flows are inconsistent; there are conflicting arguments in theory on if the volatility risk of exchange rate impedes international trade. With regard to earlier studies based on theoretical models can be stated that there is still no clear evidence of a specific effect of exchange rate volatility on international trade. It is usually assumed that if there is any impact of exchange rate volatility on trade flows, it is usually not significant (Taglioni, 2002). Researches since 1970 show a quite different refute of negative impact of exchange rate volatility on foreign trade hypothesis, only some researches prove hypothesis of negative effect of exchange rate volatility on international trade (Ozturk, 2006). Averagely proven effects are not sufficiently robust so that they can be generalized across all countries. Coric and Pugh (2010) accurately summarized this situation with argues that exchange rate volatility does have a negative impact on trade flows, but this effect is highly conditional. Empirical studies tend to confirm the significant effects of exchange rates in the case of trade with close neighbours and especially between integrated economies (Auboin and Ruta, 2011). Foreign trade of small economies is relatively more affected by exchange rate volatility than trade flows of large economies, and are more dependent on the sector and business partners (OECD, 2012). Hutchet-Bourdon and Korinek (2011) argue that there is no consensus regarding the impact of exchange rates and exchange rate volatility on the trade volume. Many empirical studies have failed to establish any significant link between exchange rate volatility and the volume of international trade (Aristotelous 2001, Assery and Peel 1991, Gagnon 1993, Tenreyro, 2004).

Methodology and Model Estimation The sample consists of four CEE countries: Czech Republic, Poland, Hungary and Slovakia with floating exchange rate regime (Slovakia only until 2008). The data sources are European Central Bank and Eurostat statistical databases. Analysis period starts in 1999 and ends in 2013, respectively 2008 in case of Slovakia. The empirical quantification is conducted using monthly, seasonally adjusted data of export of CEEC to euro area countries between period from 1999M01 to 2013M03. Bilateral exports of country to European Union are considered. A nominal export flow, denominated in eur, is the product of the export quantity Q and the export price in eur. The monthly average of bilateral nominal exchange rate, defined as currency of country price of euro data is used; volatility is estimated by G/ARCH models. For exchange rate volatility estimation, I chose the most suitable model according Akaike information criteria. Model TARCH was used for volatility modelling in case of Hungary, ARCH in case of Slovakia, and GARCH in case of Czech Republic and Poland. As external competitiveness variable, real effective exchange rate (trading partners: euro area countries) was used as a proxy variable. Lack of monthly data for GDP or income leads to application of industrial production index of euro area as a proxy variable for the economic condition of euro area. Following empirical studies in this area, I employ traditional export demand function. The model is specified as follows: EXt = β0 + β1 IPIt + β2 Pt + β3 Vt + ut

where EXt is export volume, IPIt is a logarithm measure of industrial production index in euro area countries, Pt denotes relative price, an indicator of external competitiveness expressed as logarithm of real exchange rate and Vt is the measure of exchange rate volatility, is error term.

Unit Root Test Before estimating the cointegration parameters, the order of integration of each series should be examined. The order of integration of individual time series is determined using the augmented Dickey-Fuller (ADF) test recommended by Engle and Granger (1987). The ADF test statistics for a variable yt is given by the t statistics on the estimated coefficient α, in the following regression:

where is variable, denotes the first difference, t is time trend variable, approximates white noise process and k is number of lags of yt, incorporated for residual

autocorrelation possibility. The specification is then used to test: H0: = 0, H1: < 0, nonrejection of null hypothesis implies that the series is non-stationary. Regarding the nature of time series, tests with trend and/or constants or with no supplement were used. Preliminary unit root tests applied to the data suggest that most series are non-stationary. Stationarity was in all cases proved for the first differences in all time series, thus all time series are considered as integrated of order one I(1).

Testing for cointegration Vector autoregression (VAR) models define stationary relationships and common stochastic trends. These models are widely used for monitoring interactions between individual variables and to test common trend of several time series. In these models, all variables are treated as endogenous. Unlike the AR models, VAR models recognize links between individual variables at the same time or with a time lag. The general relation of the VAR model can be written in the following form:

xt  c  1 xt 1  2 xt 2  ...  P xt P  Bxt  et where

(5)

xt is vector of variables at time t, c and B are m x 1 vector of constant and trend

e  coefficients, i is a time-invariant m x m matrix and t is a m x 1 vector of error terms with zero vector mean values, m is number of equations and endogenous variables, p is maximum lag length. This model is rewritten in the error-correction form:

where and are m x m matrices. When estimating VAR model, stationary data must be used because the advantageous features of the VAR models are valid only in the case of stationarity. Following crucial aspect of VAR model estimation is a selection of optimal lag length. For this purpose, Schwartz Information Criterion was used. The optimum lag length is four for variables of Czech Republic, Poland, Hungary and two for Slovakia. The idea is that although the variables are non-stationary, linear combination of them may be stationary, given that all variables are integrated of the same order. To confirm if variables are cointegrated or not, Johansen cointegration test was run. For the number of cointegrating vectors presence, the test was used with linear deterministic trend, assuming that there is interception but no trend in CE and VAR test. The null hypothesis of r=0 is rejected in favour of the alternative hypothesis r ≥ 1. According to trace statistics similar conclusion in each equation is obtained. The variables are cointegrated, thus are in long-run association. In sum, there is a presence of one cointegrationg vector for each country in the sample, meaning there

is one cointegration relationship among export volumes, industrial production index, real exchange rates and exchange rate volatility in all CEE countries. See Tab. I for cointegration test results, where r denotes the number of cointegrating vectors.

I: Johansen Cointegration Test Results Trace Statistics Null: Czech Republic Slovakia Poland Hungary

r=0 64.1403 (0.0007) 64.2619 (0.0007) 53.0213 (0.0151) 48.9513 (0.0393)

r≤1 23.8552 (0.2066) 25.4228 (0.1469) 27.9151 (0.0812) 28.3392 (0.0729)

r≤2 9.3312 (0.3356) 4.4839 (0.8609) 11.3116 (0.1930) 15.1768 (0.0558)

r≤3 3.5591 (0.0592) 0.3049 (0.5808) 2.6750 (0.1019) 3.0761 (0.0794)

Notes: r denotes the number of cointegrating vectors. p-value in parentheses. Source: own elaboration

Vector Error-Correction Model VECM can be developed as:

where is lagged error correction term which captures the adjustment toward the long-run equilibrium. The coefficient α denotes the proportion of disequilibrium in exports in one period corrected in next period. The estimation results are summarized in Tab. II. The error correction term coefficient for all countries is not statistically significant; meaning the conclusion about the domestic export adjustment cannot be confirmed. Besides the error term, interesting is how the current change of home export responses to exchange rate volatility in particular countries. The Tab. II shows that the coefficients of volatility are insignificant. Their values in all countries are all under 0.05. This small short run effect (∆Vt-1) shows that increases in exchange rate volatility could depress the trade flows in Poland and stimulate trade flows in Hungary, Slovakia and Czech Republic while coefficient ∆Vt-2 would indicate stimulation of trade flows in Slovakia and Czech Republic and reduction in Hungary and Poland. The effect seems to be stronger in case of Poland. However, empirical results indicate no significant short-run relationship between export and exchange rate volatility. The change in exports is positively related to changes in foreign income which is consistent with theoretical considerations. This relation was not found in case of export and real exchange rate.

II: Vector Error Correction Model Results

∆Vt-4 AIC

-3.1324

-3.4492

-3.1405

-2.8773

0.45

0.45

0.48

0.40

Constant ECTt-1 ∆Xt-1 ∆Xt-2 ∆Xt-3 ∆Xt-4 ∆IPIt-1 ∆IPIt-2 ∆IPIt-3 ∆IPIt-4 ∆Pt-1 ∆Pt-2 ∆Pt-3 ∆Pt-4 ∆Vt-1 ∆Vt-2 ∆Vt-3

R2

Poland 0.0194 [4.5273] 0.0017 [1.4416] -0.7183 [-8.1327] -0.3196 [-2.7769] -0.0058 [-0.0477] -0.1511 [-1.5482] 0.8382 [2.1817] 1.0855 [2.8601] 0.9207 [2.4270] -0.1961 [-0.5113] 0.0419 [0.2656] -0.0940 [-0.5496] 0.0443 [0.2581] -0.0849 [-0.5212] -0.0337 [-1.5463] -0.0044 [-0.2082] -0.0084 [-0.4136] -0.0305 [-1.6288]

Hungary 0.0153 [3.4702] -0.0028 [-0.039] -0.8336 [-8.0405] -0.6087 [-5.2900] -0.1937 [-1.7231] -0.1841 [-2.0708] 1.8881 [4.2066] 2.3876 [5.3388] 0.5273 [1.1432] -0.9672 [-2.0063] -0.1015 [-0.4956] -0.1129 [-0.5358] 0.0557 [0.2614] -0.0768 [-0.3671] 0.0257 [0.3846] -0.0195 [-0.2969] 0.0259 [0.4003] 0.0444 [0.6935]

Slovakia

Czech Rep. 0.0168 [3.6562] 0.01055 [0.3812] -0.6741 [-6.8943] -0.2507 [-2.0720] 0.1422 [1.1878] -0.0661 [-0.7092] 0.9649 [2.2339] 1.6642 [3.9962] 1.0169 [2.3551] -0.5769 [-1.2718] -0.0713 [-0.2309] -0.5315 [-1.7331] -0.4467 [-1.4813] 0.0981 [0.3272] 0.0015 [0.0772] 0.0000 [0.0036] -0.0003 [-0.0163] 0.0043 [0.2179]

0.0190 [3.1162] 0.0274 [1.1992] -0.7183 [-7.4521] -0.2602 [-2.6881]

2.2296 [4.0001] 1.9387 [3.1760]

0.2061 [0.5939] -0.1417 [-0.4039]

0.0003 [0.0095] 0.0302 [1.1334]

Notes: t statistics is in []. Source: own elaboration

ARDL approach The basic advantage of Autoregressive Distributed Lag (ARDL) method is ability to identify long-term relationship between variables. This enables to avoid the Johansen method problem, which occurs when there is more than one cointegrating vector. ARDL can be used regardless of whether the regressors character is I (1) or I (0). This is especially valuable for small samples, where the thickness of traditional test unit root is very small. Generally, the model is written as ADL (m, n, p), model ADL (1,1) with lag of one period can be writen as follows:

where and are stationary variables, is white noise. To capture the dynamics, the empirical framework is specified as an autoregressive distributed lag process combined with G/ARCH models used to estimate exchange rate volatility risk. To assess the impact of nominal exchange rate volatility on export volume, following equation is estimated: EXt = α0 + α1 EXt-1 +β0IPIt+ β1 IPIt-1+ γ0 Pt +γ1 Pt-1 +δ0Vt+ δ1Vt-1+εt

III: ARDL Model Results

Constant EXt-1 IPIt IPIt-1 Pt Pt-1 Vt Vt-1

Czech Rep. 0.0102 (0.0080) -0.4760 (0.0000) 1.2138 (0.0011) 0.4401 (0.2386) 1.2551 (0.0000) -0.5608 (0.0398) -0.0376 (0.0461) 0.0090 (0.6322)

Poland 0.0132 (0.0001) -0.5376 (0.0000) 1.4029 (0.0000) 0.5896 (0.0792) 0.20391 (0.1678) -0.2605 (0.0923) -0.0157 (0.3925) -0.0163 (0.3668)

Hungary 0.0086 (0.0389) -0.4772 (0.0000) 1.9809 (0.0000) 0.9680 (0.0240) 0.1759 (0.4029) -0.3909 (0.0742) 0.0513 (0.4496) -0.0561 (0.4019)

Slovakia 0.0142 (0.0220) -0.5323 (0.0000) 1.3574 (0.0109) 1.9334 (0.0011) 0.3824 (0.2915) -0.3055 (0.3943) -0.0177 (0.4770) 0.0133 (0.5817)

Notes: p-value in parentheses. Source: own elaboration According to ARDL model (see Tab. III), Czech Republic’s exports are negatively influenced by exchange rate volatility (Vt) in a very short period, coefficient is statistically significant. Results for Poland, Slovakia and Hungary show similar results as in the case of VECM model. The coefficients are not statistically significant. In all cases the constant was significant, as well as the previous development of exports, and industrial production index (except for Czech Republic).

Conclusion The Central and Eastern European EU Member States in the beginning of 1990’s relied on pegging the exchange rate to a stable currency to achieve macroeconomic stabilization. In early 2000’s, these countries moved towards more flexible regimes. This study examined the impact of exchange rate volatility on four open CEE countries’ exports to euro area using monthly data from January 1999 to March 2013. The impact is analyzed by employing errorcorrection model and Autoregressive distributed lag model. In measuring exchange rate volatility of the currencies against euro, this study employed the GARCH methodology,

modelling variance for each currency using the most suitable G/ARCH model. The impact of exchange rate volatility turns out to be ambiguous, meaning in the short-run economies that aim to stabilize the exchange rate are possibly not likely to increase the volume of export in CEE countries. Euro area economy performance has significant effect on exports, thus export performance in CEEC is highly depending on non-domestic factors and development in euro area. In case of Slovakia, adopting euro as a single currency brought other advantages. For Slovakia as a small open economy is equilibrium of export and import important with regards to influence of national economy, euro adoption meant reduction of uncertainty about exchange rates development in international trade. The further research is required since the relationship between exchange rate volatility and exports is not still solved as shown in previous researches. For Hungary, Czech Republic and Poland is exchange rate a tool for dealing with exogenous shocks. If further research is consistent with non-significant influence of exchange rate volatility on export, then fast adoption of euro in these countries is not necessary.

References ABEYSINGHE, T., YEAK, T. J., 1998: Exchange rate appreciation and export competitiveness. The case of Singapore. Applied Economics, 30(1), pp. 51-55. ARISTOTELOUS, K. (2001). Exchange-Rate Volatility, Exchange-Rate Regime, and Trade Volume: Evidence from the UK-US Export Function (1889-1999). Economics Letters, 72(1), pp. 87-94. ASSERY, A., PEEL, D. A., 1991: The Effects of Exchange Rate Volatility on Exports. Economics Letters, 37, pp. 173-177. AUBOIN, M., RUTA, M., 2011: The Relationship between Exchange Rates and International Trade: A Review of Economic Literature. WTO Working Papers, 17, 2011. [cit. 2013-09-24]. Cited from http://www.wto.org/english/res_e/reser_e/ersd201117_e.pdf CORIC, B., PUGH, G., 2010: The Effects of Exchange Rate Variability on International Trade: A Meta-Regression Analysis. Applied Economics, 42(20), pp. 2631-2644. GAGNON, J., 1993: Exchange Rate Variability and the Level of International Trade. Journal of International Economics, 25, pp. 269-287. HADDAD, M., PANCARO, C., 2010: Can Real Exchange Rate Undervaluation Boost Exports and Growth in Developing Countries? Yes, But Not for Long. Economic Premise, 20, pp. 1-5. CLARK, P., TAMIRISA, N., et al., 2004: Exchange Rate Volatility and Trade Flows - Some New Evidences. IMF Occasional Paper, No. 235. [cit. 2013-09-23]. Cited from http://www.imf.org/external/np/res/exrate/2004/eng/051904.pdf HUTCHET-BOURDON, M., KORINEK, J., 2011: To What Extent Do Exchange Rates and Their Volatility Affect Trade? The Case of Two Small Open Economies, China and New Zealand. OECD Working Papers. OZTURK, I., 2006: Exchange Rate Volatility and Trade: A Literature Survey. International Journal of Applied Econometrics and Quantitative Studies, 3(1), pp. 85-102.

STAIGER, R. W., SYKES, A. O., 2010: ´Currency Manipulation´ and World Trade. World Trade Review, 9(4), pp. 583-627. TAGLIONI, D., 2002: Exchange Rate Volatility as a Barrier to Trade: New Methodologies and Recent Evidences. Economie Internationale, 1(2), pp. 227-259.

Suggest Documents