The Effects of the Real Exchange Rate Volatility and Misalignments on Foreign Trade Flows in Uzbekistan

Olimov, Ulugbek; Sirajiddinov, Nishanbay Working Paper The Effects of the Real Exchange Rate Volatility and Misalignments on Foreign Trade Flows in Uzbekistan Economics Discussion Papers, No. 2008-29 Provided in Cooperation with: Kiel Institute for the World Economy – Leibniz Center for Research on Global Economic Challenges Suggested Citation: Olimov, Ulugbek; Sirajiddinov, Nishanbay (2008) : The Effects of the Real Exchange Rate Volatility and Misalignments on Foreign Trade Flows in Uzbekistan, Economics Discussion Papers, No. 2008-29, Kiel Institute for the World Economy (IfW), Kiel This Version is available at: https://hdl.handle.net/10419/27465 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. http://creativecommons.org/licenses/by-nc/2.0/de/deed.en D iscussion Papers Discussion Paper 2008-29 October 1, 2008 The Effects of the Real Exchange Rate Volatility and Misalignments on Foreign Trade Flows in Uzbekistan Ulugbek Olimov and Nishanbay Sirajiddinov Center for Economic Research, Tashkent, Uzbekistan Abstract: This study documents a quantitative analysis of exchange rate volatilities and misalignment in Uzbekistan for the period of 1994q3–2005q2. The results suggest that the real exchange rate volatility and misalignment have depressing effects on the volume of trade, mainly exports in Uzbekistan. The Government’s currency rationing policy was lessening the volatility proving that the policy-induced changes in exchange rate has a stabilizing effect on trade flows. The implied elasticity for the most significant real exchange rate volatility coefficient is –0.20. Using a two-step Engle-Granger technique import demand and export supply price elasticities are computed. The results are consistent with the predictions from a number of previous studies, and in particular, the estimated exports price elasticity for Uzbek economy ranges from 1.65 to 1.84, while import demand price elasticity is between –0.78 and –0.83. At the same time, relatively lower elasticity during “the currency rationing” period indicate that large devaluations, most likely, did not generate the expected improvements in the overall export performance. Submitted as Policy Paper JEL: C32, F41 Keywords: Real exchange rate; volatility; misalignment; trade flows; Uzbekistan Correspondence: Ulugbek Olimov, e-mail: ulugbek.o[email protected]; Nishanbay Sirajiddinov, e-mail: [email protected] This research paper was supported by an individual grant No. 06-070 from the Economics Education and Research Consortium, Inc. (EERC), with funds provided by the Eurasia Foundation (with funding from the US Agency for International Development), the Open Society Institute/Soros Foundations, the World Bank, the Global Development Network, the Carnegie Corporation, the Government of Finland, and the Government of Sweden. www.economics-ejournal.org/economics/discussionpapers © Author(s) 2008. This work is licensed under a Creative Commons License - Attribution-NonCommercial 2.0 Germany Contents 1 Introduction 3 2 Exchange rate policy in Uzbekistan: a brief overview 5 3 Related literature 8 3.1 Exchange rate and trade flows . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.2 Volatility, misalignment and trade flows . . . . . . . . . . . . . . . . . . . . 9 4 Exchange rate volatility and misalignment 11 4.1 Measuring the real exchange rate volatility . . . . . . . . . . . . . . . . . 11 4.2 Measuring misalignment with a single-equation approach . . . . . . . . . . 13 4.3 Exchange rate volatility and currency rationing policy . . . . . . . . . . . . 21 5 Exchange rate and trade flows: key findings 22 5.1 The real exchange rate variability and import flows . . . . . . . . . . . . . 23 5.2 The real exchange rate variability and export flows . . . . . . . . . . . . . 25 6 Conclusions 28 References 31 Tables 1 Real Exchange Rate Volatility: Sample Means . . . . . . . . . . . . . . . . 10 2 RER Volatility Measures: Summary Statistics . . . . . . . . . . . . . . . . 12 3 Stationarity Statistics: Unit Root Tests . . . . . . . . . . . . . . . . . . . . 16 4 Results of Cointegration Analysis . . . . . . . . . . . . . . . . . . . . . . . 17 5 Long Run Parameter Estimates: Two-Step Engle-Granger Method . . . . . 18 6 Observed, Equilibrium Real Exchange Rate and Misalignment . . . . . . . 19 7 Results of Cointegration and Unit Root Tests . . . . . . . . . . . . . . . . 23 8 Import Demand Equations: Error Correction Model Results . . . . . . . . 24 9 Comparative Table: Price Elasticity of Aggregate Import Demand . . . . . 25 10 Export Supply Equations: Error Correction Model Results . . . . . . . . . 26 11 Comparative Table: Price Elasticity of Aggregate Export Supply . . . . . . 27 Figures 1 The RER Dynamics and Policy Episodes in Uzbekistan, 1994–2005 . . . . 6 2 The Real Exchange Rate, Trade Flows, and the Terms of Trade, 1994–2005. 7 3 The RER Volatility Measures: Standard Deviation, ACRH and GARCH. . 13 4 Misalignment, Observed and Equilibrium Real Exchange Rates . . . . . . . 20 5 Exchange Rate Volatility vs. Currency Rationing Policy . . . . . . . . . . 21 6 Uzbekistan: Actual vs. Predicted Trade Flows . . . . . . . . . . . . . . . . 28 “I find it difficult to imagine a world in which it would not be important to estimate exchange rate misalignments.” John Williamson 1 Introduction The real exchange rate is one of the essential economic indicators of economy’s international competitiveness, and therefore, has a strong influence on country’s foreign trade developments. In particular, the impact of the real exchange rate developments on foreign trade has been a topic of discussions in both developed and developing economies. The relationship between exchange rate movements and trade flows has been studied in a large number of theoretical and empirical papers. It is commonly accepted that the movements of the real exchange rate have a permanent effect on exports and imports. There are several studies that empirically tested this hypothesis using data mostly from developed and developing countries.1However, there are several studies that failed to find evidence of an effect of the real effective exchange rate on trade flows.2. In the light of recent developments in economic policy in Uzbekistan it becomes quite important to examine the pros and cons of different exchange rate policy strategies. Particularly, in 1996, after low export revenues, the Government took measures to restrict convertibility of the Soum (national currency – UZS) by reducing its quantity that could be officially converted into foreign currency. Since that time there is existed a multiple exchange rate regime in Uzbekistan, which included the official, the commercial bank and parallel market exchange rates. The Government adopted protectionist measures as part of the import-substitution strategy. In this context, an overvalued official exchange rate has been used to tax exporting sectors (cotton and gold production) in order to subsidize imports of capital and priority consumer goods. As a result, the illegal curb market for foreign exchange rate gained in importance. Staring from October 2003 Uzbekistan has fulfilled its currency convertibility obligations to the IMF and joined the Article VIII of the IMF. The above changes in economic policy and realities of the current transition period, among other things, put the issue of formulation of an effective exchange rate policy on the top of policy agenda and requires comprehensive analysis and modelling of exchange 1For instance, Mordecki (2000) concludes that during the period of 1990–1998 the Uruguayan external sector was permanently affected by the changes in exchange rate. 2Rose (1991) and Lahreche-Revil and Benassy-Quere (2001) were not able to evidence an influence of exchange rate uncertainty on trade in the G-7 countries and in China, respectively. 3 rate behavior in new conditions. In particular, the movements of the real exchange rate should be carefully analyzed and systematically monitored. In this regard, this paper investigate the effect of exchange rate variability on the Uzbek imports from and exports to other countries.3 It is often believed that the real exchange rate is overvalued in those economies where restrictive foreign trade policy has been implemented. In this respect, this paper attempts to measure the degree of a possible misalignment in the real exchange rate and to analyze its impact on foreign trade flows in Uzbekistan. Recent empirical studies suggest that the analysis of exchange rate misalignments plays a key role in designing of policy strategies and may serve as a potential predictor of currency crisis in emerging economies (Kaminsky, Lizondo, and Reinhart 1998). Most studies find evidence for strong negative effects on a country’s economic growth rate (for example, Cottani, Cavallo, and Khan (1990), Ghura and Grennes (1993), Yotopoulos (1996)). At the same time, there is a number of cross-country studies investigating the relationship between distortion in real exchange rate and economic growth. They conclude that trade liberalization and devaluation of the real exchange rate could improve export performance and trigger export-led growth (Dollar (1992), Popov and Polterovich (2004)). The results of this study would assist in assessment of the degree of impact of the exchange rate and show up the possible developments in the case of different policy instruments (including devaluation of exchange rate, tariffs and import protectionism etc.) which also will be useful in designing of export promotion policy.4 The estimation results would give an idea for policy makers about the level of the real exchange misalignment and serve as a basis for policy-oriented recommendations related to the modelling and choice of the appropriate the real exchange misalignment. For instance, foreign trade liberalization and greater exchange rate flexibility could increase of both trade flows and exchange rate volatility. Therefore, appropriate estimations, and consequently policies are needed to avoid the underlying causes of unpredictable and damaging movements in exchange rates. In addition, to our knowledge there is no study on the subject of this paper that documented the impact of real exchange rate variabilities on foreign trade flows in Uzbekistan using econometric techniques. Therefore, in this paper we have used a comprehensive ap3Despite ongoing developments, some recent studies characterize Uzbekistan as a slow reformer, lagging behind in trade and foreign exchange liberalization. E.g. the EBRD Transition Report 2006 and the World Banks Doing Business 2007 survey are among others. 4The results reported in this paper have been generated using Stata, Ox (available free for research purposes) and R codes (open source). 4 proach to both methodology and interpretation of obtained results in terms of bringing up details. The plan of the paper is as follows. Section 2 begins with review of the exchange policy episodes for the period of 1993–2005 in Uzbekistan. Section 3 briefly reviews the relevant (selected) literature on empirical evidence of exchange rate volatility and misalignments in both developed and developing countries. In addition, various estimates of exchange rate volatility and misalignment measures for Uzbekistan are reported. Then, Section 4 lays out the theoretical framework. Following the existing literature briefly discussed in Section 3 we estimate several proxies of real exchange rate volatility and misalignment for Uzbekistan. Section 5 contains the empirical results. Finally, Section 6 presents conclusions. 2 Exchange rate policy in Uzbekistan: a brief overview In Uzbekistan the exchange rate policy can be divided into three periods: the period between 1993 to 1996, late 1996 to 2000, and the post 2000 period. The first period starts when in November 1993 the Government of Uzbekistan introduced a temporary (intermediate) currency “soum-coupon” at the rate of 1:1 to the Russian rouble. In 1993–1994, the Government took several important measures to develop foreign exchange rate market in the country. In July 1994 a full-value national currency “soum” has been introduced at the rate of 1:1000 to the soum-coupon and 7:1 to the U.S. dollars. Next three months exchange rate of UZS stabilized at level of 11–12 UZS/USD. During the second period, to protect the domestic producers from foreign competition a strategy of import substitution was implemented. In late 1996, after low export revenues, the Government took measures to restrict convertibility of UZS by reducing its quantity that could be officially converted into foreign currency (currency rationing policy). Since that time there is existed a multiple exchange rate regime, which included the official (traded through the Republican Currency Exchange – RCE), the commercial bank (traded through the over-the-counter – OTC – currency market) and parallel market (curb) exchange rates. The Government regularly intervened in the official and commercial markets by controlling the supply and demand for foreign exchange. An overvalued official exchange rate has been used to tax exporting sectors (cotton and gold production) in order to subsidize imports of capital and priority consumer goods. As a result, the illegal curb market for foreign exchange rate gained in importance (IMF 2000). The official rate is used for surrendering the proceeds from centralized exports, 5 Figure 1: The RER Dynamics and Policy Episodes in Uzbekistan, 1994–2005 Introduction of currency rationing Unification of official and OTC rates (60% dev.) OTC market expansion Unification of CBU & OTC rates (60% dev.) Unification of official and market rates 60 80 100 120 140 160 180 200 220 60 80 100 120 140 160 180 200 220 1994q3 1996q1 1997q3 1999q1 2000q3 2002q1 2003q3 2005q1 official market Note: An upward movement is an appreciation of the RER. The time interval between dashed vertical lines denotes the Government’s currency rationing policy period. Source: Authors’ calculations. imports of certain capital goods and consumer goods, government purchases, and servicing of government and government-guaranteed debt. As a result of absence of free foreign exchange market and of coordination between exchange rate, monetary and fiscal policies, the spread between official rate and parallel market rate is gradually increased, and for instance, had widened up to 440% (!) by the second quarter of 2000. Besides, the restrictive foreign exchange and trade regime has led to both import and export declines since 1996 (Figure 2-a). Indeed, large fall in exports prices, mainly world cotton and gold prices, also had a negative impact on export earnings (Figure 2-b). The currency rationing policy had also other negative consequences such as strong distortions of the relative prices; restraint of inflow of direct foreign investments; subsidizing of importers due to exporters; monopolization of the market; and absence of macroeconomic stability. The third period starts in 2000, when the Government announced gradual liberalization of the foreign exchange and trade regimes. The policy targets were straightforward: reduce the spread between the official and parallel market exchange rates, establishing a realistic and single market rate, liberalizing both access and sale of foreign exchange and promote 6 Figure 2: The Real Exchange Rate, Trade Flows, and the Terms of Trade, 1994–2005. 100 150 200 250 300 350 400 450 100 150 200 250 300 350 400 450 1994q3 1996q1 1997q3 1999q1 2000q3 2002q1 2003q3 2005q1 exports imports (a) Exports & Imports, 1994q3=100 70 80 90 100 110 120 130 140 150 160 170 180 70 80 90 100 110 120 130 140 150 160 170 180 1994q3 1996q1 1997q3 1999q1 2000q3 2002q1 2003q3 2005q1 WRER ToT (b) The RER & the Terms of Trade Notes: WRER is defined as weighted average official and parallel market real exchange rates i.e. WRER= 0.65 ∗RERO+ 0.35 ∗RERM. The Terms of Trade is proxied by cotton and gold prices and equal to 0.75 ∗COT + 0.25 ∗GOL. An upward movement is an appreciation of the RER and/or WRER. Time interval between dashed vertical lines denotes the Government’s currency rationing policy period: 1996q3 – 2003q4. Sources: computed from the IMF/IFS data. Authors’ calculations. exports. Particularly, the official rate is allowed to depreciate (against the U.S. dollar) by more than 60% in the first half of 2000, and the spread between the official and parallel market rates has been reduced to about 2.5 times compared to about 3.5 times during certain periods of 1999. For the purpose of expansion of the over-the-counter exchange market all currency operations, but external debt servicing, have been transferred to the OTC market. As a result of the undertaken measures in May 2000 the RCE (official) and OTC exchange rates were unified. In November 2001, the Central Bank’s official exchange rate devalued by about 60% against the dollar, and it was unified with the OTC rate (Figure 1). At the end of 2001, the U.S. dollar was trading at 1497 UZS on the parallel market compared with the official rate of 693 UZS. During 2002 the official exchange rate was devalued in steps by about 30% against the U.S. dollar. At the same time, the parallel market exchange rate appreciated by 12% by year-end; however, the government failed to achieve the 20% target and the spread was equal to 38%. Taken measures have allowed to unify official and the parallel market exchange rates. Staring from 15 October 2003 Uzbekistan has fulfilled its currency-convertibility obligations 7 to the IMF and joined the Article VIII of the IMF. Refusal from currency rationing and achievement of convertibility of national currency have considerably strengthened an export orientation of economic policy, as became one of the major factors of growth of foreign trade. 3 Related literature 3.1 Exchange rate and trade flows It is commonly accepted that the movements of the real exchange rate have a permanent effect on exports and imports.5There are several studies that empirically tested this hypothesis using data mostly from developed and developing countries. For instance, Mordecki (2000) concludes that during the period of 1990-1998 the Uruguayan external sector was permanently affected by the changes in exchange rate. Several studies examined the role of exchange rate policies in export promoting and the effects of exchange rate volatility on exports. McKenzie (1999) provides a comprehensive survey of the existing literature and concludes that the results depends on the estimation technique of volatility. Moreover, most authors argue that short-run volatility has a transitory effect, and that long-run volatility has more impact on trade flows (Sapir, Sekkat, and Weber (1994), Sekkat (1998), Rose (2000) among others). Another interesting study documented strong evidence that Sri Lanka’s exports to the developed countries under investigation were adversely affected by the increased volatility in bilateral real exchange rates during 1978–96 (Weliwita, Ekanayake, and Tsujii 1999). Sekkat and Vaoudakis (1999) made similar conclusion when they analyzed the impact of exchange rate policies on export performance in four North African countries.6They also found negative impact of the exchange rate volatility on trade flows. Lane and Milesi-Ferretti (2001) provides additional evidence that there is a negative long-run association between the trade balance and the real exchange rate, and the relative price of non-tradables is an important channel linking the trade balance and the real exchange rate. Using monthly data Pickard (2003) examines trade flows of certain steel products between Canada, Mexico and the United States. The main conclusion of this 5However, Rose (1991) failed to find evidence of an effect of the real effective exchange rate on trade flows for the G-7 countries. Lahreche-Revil and Benassy-Quere (2001) were also unable to evidence an influence of exchange-rate uncertainty on trade in China. 6They considered Algeria, Egypt, Morocco and Tunisia and the period of 1970–92. 8 for trade restrictions and/or controls. Certainly, more liberal the trade regime the more total amount of foreign trade. However, its impact on real exchange rate depends on the nature (“direction”) of this increase, and therefore the expected sign is ambiguous. For instance, if the current account worsened, then the real exchange rate would be depreciated. •gcons = share of government consumption in GDP. This is a proxy for government demand for nontradables.In the case of increase in government consumption will result in appreciation of the real exchange rate. •ishare = share of investment in GDP. The expected sign is ambiguous. Because it heavily depends on the relative factor intensities between sectors of economy. •gdp = per capita GDP. It is proxy for domestic productivity. •d96 = a policy change dummy variable that takes the value of one for the period of 1996q3–2003q4 and zero otherwise. Since the Governments currency rationing policy has been directed to tax exporting sectors (mainly, cotton and gold production) in order to subsidize imports of capital and priority consumer goods, this shock should appreciate the real exchange rate. The estimation strategy consists of three steps. First step involves estimating the longrun parameters. We employ the Johansen-Juselius cointegration test in order to investigate the existence of a long-run relationship between the real exchange rate and the selected explanatory variables (determining the order of integration). Another econometric issue we attend is nonstationarity of variables. In this case standard regression techniques such as the Ordinary Least Squares are not useful and could lead to “spurious” results. For that reason, we also report the results of nonstationarity tests i.e. the Augmented Dickey-Fuller and the Philip-Perron unit roots tests. In the second step, the proposed fundamentals are decomposed into permanent and transitory components. Then the transitory effects are set to zero and actual values of the fundamentals are inserted into the estimated equation. In the final step the real exchange rate misalignments will be estimated as the residuals of Equation 3. Table 3 shows the results of the Augmented Dickey-Fuller and Phillips-Perron unit-root tests for all variables. Both tests indicate that the real exchange rate and terms of trade are seemingly nonstationarity variables. At the same time, all other variables appear to be 15 Table 3: Stationarity Statistics: Unit Root Tests Without TT With TT Without TT With TT ADF PP ADF PP ADF PP ADF PP rert-1.11 -1.38 -1.87 -1.90 ∆rer -4.91 -4.97 -5.31 -5.35 tott-1.38 -1.56 -1.28 -1.60 ∆tott-4.59 -4.49 -4.57 -4.45 opent-1.76 -2.38 -3.15 -2.84 ∆opent-8.60 -11.11 -8.80 -10.42 gconst-2.28 -2.98 -3.61 -3.51 ∆gconst-7.44 -8.59 -7.52 -9.14 isharet-2.57 -2.06 -5.73 -5.73 ∆isharet-7.72 -14.30 -7.62 -14.05 gdpct-2.18 -3.82 -4.61 -4.69 ∆gdpct-7.35 -11.76 -7.40 -15.71 Notes: TT – time trend. ADF and PP refer to augmented Dickey-Fuller and Phillips-Perron tests for a unit root, respectively. The number of observations is 44. Source: Authors’ calculations. trend-stationary. The unit root test results also confirm that all variables are stationary in first difference i.e. I(1). It means that we could employ simple first differencing technique to remove the nonstationarity problem. However, this could result in discarding useful information about relationship among the variables of interest, for example, in our case the relationship between the real exchange rate and its fundamentals. Therefore we first test for the presence of cointegration between rertand Ft. Table 4 reports the estimated cointegrating relationships, and also shows the presence of I(1) cointegrating vectors. Using the λ-max statistic (column 2), we test the null hypothesis that r= 0 against the alternative that r≤1. Our test statistic of 113.44 exceeds the critical value of 45.28 which leads to the rejection of the hypothesis of no cointegrating relationship. The trace statistic of 176.86 also exceeds its corresponding critical value of 124.40 which is consistent with the result using the λ-max statistic. Next we test the null hypothesis of at most one cointegration vector (r= 1). In this case, however, the λ-max statistic of 28.30 is smaller than the critical value of 39.37 and we cannot reject the null hypothesis. The trace test leads to the same conclusion. Therefore, regardless of which statistic is used, we cannot reject the hypothesis that we have one cointegrating vector. We now move on estimating the long-run parameters using the nonstationary panel techniques, in particular, the Engle-Granger two-step method (based on Equation 3). Ta16 Table 4: Results of Cointegration Analysis 5 percent critical values 10 percent critical values λ-max trace λ-max trace λ-max trace With the policy change dummy variable r= 0 113.44 176.86 45.28 124.40 42.32 118.50 r≤1 28.30 63.41 39.37 94.15 36.76 89.48 Without the policy change dummy variable r= 0 103.54 154.78 39.37 94.15 36.76 89.48 r≤1 13.28 21.67 33.46 68.52 30.90 64.84 Note: The first row (r= 0) tests the null hypothesis of no cointegration; the second (r= 1) tests the null hypothesis of at most one cointegration vector, etc. λmax – the estimated Johansen-Juselius likelihood value. trace – trace statistics. The Osterwald-Lenum critical values (95% and 90% confidence intervals). Sample size – 44, 1994q3:2005q2. Source: Authors’ calculations. ble 5 contains long-run parameter estimation results for Uzbekistan obtained from the Engle-Granger method, including and excluding the 1996 policy change dummy variable. Unit root tests applied to the estimated residuals obtained from the first step of the Engle-Granger method indicate that there is strong evidence of cointegration in both cases (with/without the policy change dummy variable). In general the estimated long run parameters are consistent with the predictions from theory. In particular, the government consumption (gcons) has an appreciating impact on the real exchange rate. The effect is particularly strong and significant in the model estimated including the policy change dummy variable (d96) from the single-equation, suggesting that most government spending is directed toward nontradables. Investment (ishare) has a negative effect on the real exchange rate, suggesting that an increasing share of investment in GDP shifts spending toward tradable goods in the long term. This shift, other things equal, would depreciate the real exchange rate. The implied elasticity suggests that a 10% increase in the investment share in GDP depreciates the real exchange rate by at least 1.4% (2.0% in the second model).11 The estimated openness coefficient (open) is positive and significant in both models, suggesting that trade liberalization policy would appreciate the long-run equilibrium real 11According to Edwards (1989) this indicator is equal to 7% for s group of 12 developing countries. 17 Table 5: Long Run Parameter Estimates: Two-Step Engle-Granger Method The First Step The Second Step with d96 without d96 with d96 without d96 Constant -1.01 2.58 Constant -0.11 2.56** (-0.59) (1.31) (-0.16) (2.37) Adjustment speed: ut−1-0.12 -0.08 (-0.97) (-0.70) Parameters: tott0.08 -0.65* tott−1-0.13 -0.69*** (0.24) (-1.76) (-0.57) (-3.40) gdpct0.21 -0.05 gdpct−10.20* -0.01 (1.25) (-0.21) (1.76) (-0.06) opent0.27* 0.33* opent−10.34*** 0.39*** (1.94) (1.89) (3.67) (3.03 ) gconst0.13 0.50*** gconst−10.18*** 0.48*** (1.13) (2.78) (3.25) (4.59) isharet-0.12** -0.20*** isharet−1-0.14*** -0.20*** (-2.43) (-3.16) (-7.99) (-5.52) d96t0.48*** – d96t−10.38*** – (-4.98) – (5.83) – R20.68 0.52 R20.91 0.76 DW 0.88 0.92 DW 1.84 1.35 ADF -3.49 -3.70 ADF -6.19 -4.83 PP -3.55 -3.64 PP -6.20 -4.76 Notes: The numbers in parentheses are t-statistics. DW – the Durbin-Watson statistic. ADF – the augmented Dickey-Fuller test. PP – the Phillips-Perron test. Significant at the ***1% ,**5% and *10% level. The dependent variable is ∆rert.ut−1– error term. Source: Authors’ calculations. exchange rate. The terms of trade (tot) in both cases has a negative and significant (only in the model corrected for d96) impact on the real exchange rate. The positive coefficient on d96 dummy variable is consistent, and confirms that the Government’s currency rationing policy substantially appreciated the real exchange rate. Now, we can estimate the degree of misalignment, ma1,t, which is simply the percentage difference between the real and equilibrium exchange rates: ma1t= ( rert rer∗ t −1) ·100.(4) Table 6 show alternative measures of the equilibrium real exchange rate and the estimated level of the real exchange rate misalignment. The equilibrium exchange rate has 18 Table 6: Observed, Equilibrium Real Exchange Rate and Misalignment 1996q1:96q4 1999q4:2000q4 2003q1:03q4 2004q1:05q2 Observed rert154 184 103 100 Equilibrium rer∗ t126 147 94 104 misalignment I ma1t10.1 25.6 6.6 -5.5 misalignment II ma2t11.4 24.7 -1.3 -0.3 Notes: The observed RER (rert) is the official real exchange rate. ma1t= 100 ∗(rert−rer∗ t)/rer∗ t. been calculated based on the long-run parameter vectors, taken from the Engle-Granger regression in column 5 of Table 5. Equilibrium values are obtained directly from that regression. Figure 4 depicts the observed and equilibrium real exchange rates. Figure 4-b shows that misalignment patterns are consistent with different policy periods discussed in Section ??. The estimation results for 1994q3-1995q4 and 1996q1-1996q4 periods do not indicate any major overvaluation, and the real exchange rate was on average undervalued by 10%. During the Government’s currency rationing policy period (late 1996–2000), the above diagram reveals a substantial overvaluation (9% period average) and a widening gap between the actual and the estimated equilibrium real exchange rates. The period from 1999q4 to 2000q4 was , when the estimated overvaluation reached up to 38% in the first quarter of 2000 (25% period average). The Government managed to reverse substantial real overvaluation by the end of 2002. After several devaluations the real exchange rate was very close to the estimated equilibrium level. In particular, in the fourth quarter of 2002 the real exchange rate was undervalued only by 1%. The average undervaluation of the real exchange rate during the post 2002-year period was equal to 1%. In addition to the estimated degree of misalignment ma1tfrom Equation 4, we use another alternative measure of misalignment. This approach is based on using use the index of country’s relative price level (RPL), suggested by Dollar (1992), where RPLt= 100·rert·P/P∗. In other words, he used the RPL index as a measure for the real exchange rate. To estimate the relationship between national price level and endowments, Dollar (1992) 19 Figure 4: Misalignment, Observed and Equilibrium Real Exchange Rates 80 100 120 140 160 180 200 80 100 120 140 160 180 200 1994q3 1996q1 1997q3 1999q1 2000q3 2002q1 2003q3 2005q1 rer rer* (a) rertand rer∗ t, 2004=100 2001q1; 38% 1997q1; 23% 2003q2; 18% −20 −10 0 10 20 30 40 50 −20 −10 0 10 20 30 40 50 1994q3 1996q1 1997q3 1999q1 2000q3 2002q1 2003q3 2005q1 (b) ma1t: misalignment −20 −10 0 10 20 30 40 −20 −10 0 10 20 30 40 1994q3 1996q1 1997q3 1999q1 2000q3 2002q1 2003q3 2005q1 mis I mis II (c) ma1t&ma2t: misalignment measures Notes: An upward movement is an appreciation of the RER. ma1t= 100 ∗(rert−rer∗ t)/rer∗ t. used different model specifications of the following regression: RPLt=α0+α1GDPt+ α3DENSt+dt, where GDPt– per capita GDP; DENSt– population density; and dt– year dummies. We also constructed the similar equation using the Uzbek data for the period of 1996q3–2005q2. Hence the real exchange rate misalignment (ma2t) is estimated as the residuals from the the following regression:12 RPLt=α0+α1GDPt+α2GDP2 t+α3d96 + qk+trendt+ut. 12The regression number 6 in Dollar (1992), Table 2, p. 528. 20 Figure 5: Exchange Rate Volatility vs. Currency Rationing Policy 0 .1 .2 .3 .4 0 .1 .2 .3 .4 1994q3 1996q1 1997q3 1999q1 2000q3 2002q1 2003q3 2005q1 actual predicted (a) SD-based volatility 0 .2 .4 .6 .8 1 1.2 1.4 0 .2 .4 .6 .8 1 1.2 1.4 1994q3 1996q1 1997q3 1999q1 2000q3 2002q1 2003q3 2005q1 actual predicted (b) ARCH-based volatility 4.3 Exchange rate volatility and currency rationing policy As discussed earlier, it is an important issue to analyze the nature of shocks causing to real exchange rate as well as implications of the governments currency rationing policy for dynamics of the real exchange rate. For this purpose, we predict the real exchange rate at the second stage (after the cointegration analysis) using the Equation 3 and assuming the inflows of private capital and cotton prices are constant. Comparing the fluctuations of this predicted real exchange rate with the fluctuations of the actual real exchange rate would allow us to estimate to what extent government policy in particular periods was mitigating the volatility or, on the contrary, increasing this volatility. We plotted actual and predicted exchange rate volatilities in Figure 5. During the period of 1996-2002 the Governments currency rationing policy was lessening the volatility. On the contrary, gradual liberalization of the foreign exchange and trade regimes, aimed at establishing a realistic and single market rate, led to substantial increase of this volatility (the post 2002 period). 21 5 Exchange rate and trade flows: key findings In this section, we investigate to which extent, empirically, the real exchange affects foreign trade flows in Uzbekistan. In general, there is, so called, “pessimism” concerning the impact of changes in exchange rates on trade flows in developing countries. Some studies argue that due to low import and export elasticities the required changes in the real exchange rate are not able to improve the trade balance, and therefore, exchange rate adjustment policies are not effective in dealing with balance-of-payments deficits. Ghei and Pritchett (2001) examines three types of the above-mentioned pessimisms regarding the elasticity of import demand and export supply as well as the elasticity of demand for the country’s exports in world markets. They find that none of the three pessimisms is justified for developing countries. Reviewing the existing empirical literature, Ghei and Pritchett (2001) conclude that a reasonable range for the aggregate price elasticity of demand for imports is -0.7 to -0.9, and that elasticities of export supply tend to be in the range of 1.0 to 2.0. Nevertheless, in most of developing (transition) countries, in particular in Uzbekistan, exchange rate adjustment policy is considered as an essential element of trade policy. Therefore, reasonably accurate and precise estimates of import and export price elasticities can be used by policy advisors and practitioners in applied economic policy analysis, and in designing of development strategies. Following this logic, below we have tried to shed light into possible effects of the real exchange rate movements on export and import flows in Uzbekistan. In order to analyze the effects of the real exchange rate volatility and misalignments on imports and exports, consistent with the existing studies, we estimate the long-run export and import equations. In particular, the following typical log-linear specifications will be separately estimated using a two-step Engle-Granger estimation method: rm =f(rgd, rer, δi, d96) (5) rx =g(rgd, rer, δi, d96) (6) where rm and rx are real aggregate imports and real exports, respectively; rgd is domestic real GDP; rer is the real exchange rate to proxy the relative price competitiveness of commodities between countries; δi,i= 1,2,3 – the estimated measures of the real exchange rate volatility; and d96 – the 1996 policy dummy. All variables are in logs with the exception of the volatility measures. A constant, a 22 Table 7: Results of Cointegration and Unit Root Tests H0λ-max trace ˆ λˆ trace Import equation r= 0 43.82 68.76 33.46 68.52 r≤1 15.50 24.94 27.07 47.21 Export equation r= 0 45.67 70.43 33.46 68.52 r≤1 17.94 24.76 27.07 47.21 Var ADF PP Var ADF PP rmt-2.12 -2.74 ∆rmt-8.36 -9.00 rxt-4.82 -4.78 ∆rxt-12.1 -14.3 rert-1.12 -1.35 ∆rert-5.04 -5.10 rgdt-2.34 -3.95 ∆rgdt-7.07 -8.99 δ1-2.33 -2.34 ∆δ1-6.54 -6.62 δ2-1.69 -1.56 ∆δ2-6.10 -6.20 δ3-1.24 -1.66 ∆δ3-4.90 -4.95 Note: 1) Cointegration test. The first row (r= 0) tests the null hypothesis of no cointegration; the second (r= 1) tests the null hypothesis of at most one cointegration vector, etc. λ-max – the estimated Johansen-Juselius likelihood value. ˆ λand ˆ trace – the Osterwald-Lenum critical values (95% interval). Sample size – 44 quarterly observations, 1994q3:2005q2. 2) Unit roots tests. ADF and PP refer to augmented Dickey-Fuller and Phillips-Perron tests for a unit root, respectively. T= 44. trend and a set of seasonal (quarterly) dummies are included in all regressions. Regarding the expected coefficients’signs, theory predicts that the real exchange rate volatility increases uncertainty and therefore could hinder the foreign trade flows. From this point we use the “appreciation is down” convention of measuring the real exchange rate in domestic-currency terms. In this case, for example, an increase in the real exchange rate in domestic-currency terms is an depreciation. In order to examine the stationarity of the data, we conducted two unit root tests: the augmented Dickey-Fuller and the Phillips-Perron tests. The test results reported in Table 7 indicate that all variables are integrated of order one i.e. I(1), but rxtappears to be I(0) at the 1% level. Using the Johansen’s method we tested for the number of cointegrating vectors in the system. The results, reported in Table 8, also confirm the presence of I(1). 5.1 The real exchange rate variability and import flows In the presence of cointegrating vectors, it is advised to employ an error correction model developed by Engle and Granger (1987) that was applied earlier in Section 4 for estimating the long-run parameters of the real exchange rate. Table 8 shows the results for the import demand regressions with different volatility and misalignment measures. 23 Table 8: Import Demand Equations: Error Correction Model Results dep.var: Volatility dep.var: Misalignment rmtwith δ1with δ2with δ3rmtwith mis1with mis2 Constant 2.35*** 2.25*** 2.45*** Constant 2.35*** 2.39*** (8.77) (10.51) (11.09) (13.47) (15.53) Adjustment Speed Adjustment Speed rmt−1or ut−1-0.64** -0.87** -0.51** rmt−1or ut−1-0.67*** -0.58** (-4.11) (-5.98) (-3.45) (-4.65) (-3.69) rgdt0.38** 0.40*** 0.38** rgdt0.46** 0.53*** (3.02) (4.10) (3.00) (2.95) (4.65) rert-0.78** -0.83** -0.40 misi,t -0.007** -0.003 (-2.79) (-3.04) (-1.19) (-2.65) (-1.54) δi,t -0.15** -0.20*** -0.01 δ1-1.10** -0.92* (-2.33) (-5.72) (-0.32) (-2.11) (-1.86) d96 -0.22** -0.01 -0.32** d96 -0.12 -0.15* (-2.02) (-0.04) (-2.25) (-1.25) (-1.76) trend -0.05** -0.05*** -0.04** trend -0.05** -0.06*** (-2.96) (-4.10) (-2.36) (-2.97) (-4.40) R20.69 0.77 0.62 R20.70 0.68 DW 2.19 1.96 2.33 DW 2.25 2.28 Notes: The numbers in parentheses are t-statistics. DW – the Durbin-Watson statistic. Significant at the ***1%, **5% and *10% level. The dependent variable is rmt.ut−1– error term. Misalignment indicator is converted to rdtwhich indicates the estimated depreciation or appreciation required to bring the actual real exchange to the equilibrium level, and calculated as following: rdt= (rer∗ t/rert−1) ·100. Source: Authors’ calculations. As expected import demand price elasticities are negative in all three regressions, but significant only the first two models with average -0.81. The results are consistent with the predictions from a number of studies that estimate import price elasticities for both developed and developing countries (see Table 9). The real exchange rate volatility coefficients have the expected sign in all three equations13, but significant at reasonable statistical levels only in two regressions (with standard deviation and GARCH-based volatility measures). The implied elasticity suggests that a 10% increase in the volatility would cause a decrease of imports by at least 1.5% – 2.0%. Another important observation is the adjustment speed. The estimated adjustment coefficient of the cointegrating vector is for all models ranges from -0.51 to -0.87 and 13Which is consistent with the existing empirical works. E.g. Weliwita, Ekanayake, and Tsujii (1999), Sekkat and Vaoudakis (1999), Lane and Milesi-Ferretti (2001), Frankel and Wei (1993), Devlin et al. (2001) 24 References Baffes, J., I. Elbadawi, and S. O’Connell. 2001. Chapter 10 of Exchange rate misalignment: concepts and measurement for developing countries, edited by L. Hinkle and P. Montiel, 405–463. A World Bank Research Publication, Oxford University Press. Begg, D., L. Halpern, and Ch. Wyplosz. 1999. “Monetary and Exchange Rate Policies, EMU and Central and Eastern Europe.” CEPR and EastWest Institute Forum Report on the Economic Policy Initiative, vol. 5. Calderon, C. 2004. “Trade Opennes and Real Exchange Rate Volatility: Panel Data Evidence.” Central Bank of Chile, vol. WP 294. Caporale, T., and K. Doroodian. 1994. “Exchange Rate Variability and the Flow of International Trade.” Economic Letters 46:49–54. Cottani, J., D. Cavallo, and M. Khan. 1990. “Real Exchange Rate Behavior and Economic Performance in LDCs.” Economic Development and Cultural Change 39:61–76. Csajbok, A. 2003. “The Equilibrium Real Exchange Rate in Hungary: Results from Alternative Approaches.” Paper presented at the 2nd workshop on Macroeconomic Policy Research. Magyar Nemzeti Bank. Devlin, R., A. Estevadeordal, P. Giordano, J. Monteagudo, and R. Saez. 2001. “Macroeconomic Stability, Trade and Integration.” Integration and Trade 5(13):35–96. Dollar, D. 1992. “Outward-Oriented Developing Economies Really Do Grow More Rapidly.” Economic Development and Cultural Change 40(3):545–566. Edwards, S. 1989. “Real Exchange Rates, Devaluation, and Adjustment.” Cambridge, Mass.: MIT Press. Egert, B. 2003. “Assessing Equilibrium Exchange Rates in CEE Acceding Countries: Can We Have DEER with BEER without FEER? A Critical Survey of the Literature.” Focus on Transition 2:38–106. Engle, R., and C. Granger. 1987. “Co-integration and error correction: Representation, estimation and testing.” Econometrica 55:251–276. Feenstra, R., and J. Kendall. 1991. “Exchange Rate Volatility and International Prices.” National Bureau of Economic Research, Working Paper, vol. 3644. Frait, J., L. Komarek, and M. Melecky. 2005. “The Real Exchange Rate Misalignment in the Five Central European Countries.” Warewick Economic Research Papers, The Univeristy of Warwick, vol. No 739. Frankel, J., and Sh-J. Wei. 1993. “Trade Blocs and Currency Blocs.” National Bureau of Economic Research, Working Paper, vol. 4335. Ghei, N., and L. Pritchett. 2001. Chapter 11 of Exchange rate misalignment: concepts and measurement for developing countries, edited by L. Hinkle and P. Montiel, 464–496. A World Bank Research Publication, Oxford University Press. 31 Ghura, D., and T. Grennes. 1993. “The Real Exchange Rate and Macroeconomic Performance in Sub-Saharan Africa.” Journal of Development Economics 42:155–174. Hausmann, R., U. Panizza, and R. Rigobon. 2004. “The Long-Run Volatility Puzzle of the Real Exchange Rate.” NBER Working Paper, vol. WP 10751. IMF. 2000. “Republic of Uzbekistan: Recent Economic Developments.” Staff Country Report 00/36. Kaminsky, G., S. Lizondo, and C. Reinhart. 1998. “Leading Indicators of Currency Crises.” IMF Papers 45:1–48. Kim, B., and I. Korhonen. 2002. “Equilibrium Exchange Rates in Transition Countries: Evidence from Dynamic Heterogeneous Panel Models.” BOFIT Discussion Papers, vol. 15. Kroner, K., and W. Lastrapes. 1993. “The Impact of Exchange Rate Volatility on International Trade: Reduced From Estimates Using the GARCH-in-Mean Model.” International Money and Finance 12:298–318. Lahreche-Revil, A., and A. Benassy-Quere. 2001. “The Impact of Exchange Rate Strategies on Trade and Foreign Direct investment in China.” Technical Report, Paper presented at the workshop on China’s Economy, organized by the CEPII in Paris on December 12, 2001. Lane, Ph., and G. Milesi-Ferretti. 8-9 June 2001. “External Wealth, the Trade Balance, and the Real Exchange Rate.” EEA/NBER International Seminar on Macroeconomics, Dublin. Lee, J. 1999. “The Effect of Exchange Rate Volatility on Trade in Durables.” Review of International Economics 7(2):189–201. McKenzie, M. 1999. “The Impact of Exchange Rate Volatility on International Trade Flows.” Journal of Economic Surveys 13(1):71–106. Montiel, P. 1997. “Exchange Rate Policies and Macroeconomic Management in ASEAN Countries.” IMF. Mordecki, G. 2000. “Uruguayan Trade and The Real Exchange Rate: VAR Analysis 1990–1998.” Instituto de Economia, Universidad de la Republica. Nabli, M., and M. Veganzones-Varoudakis. 2002. “Exchange Rate Regime and Competitiveness of Manufactured Exports: The Case of MENA Countries.” World Bank, vol. Working Paper Series No.27. Ofair, R., and S. Collins. 1997. “Real Exchange Rate Misalignments and Growth.” NBER Working Paper, vol. WP 6174. Pickard, J. 2003. “Exchange Rate Volatility and Bilateral Trade Flows: An Analysis of U.S. demand for Certain Steel Products from Canada and Mexico.” Master’s Thesis, Faculty of the Virginia Polytechnic Institute and State University. 32 Popov, V., and V. Polterovich. 2004. “Accumulation of Foreign Exchange Reserves and Long Term Growth.” New Economic School, Moscow, Russia. Rose, A. 1991. “The Role of Exchange Rates in a Popular Model of International Trade: Does the Marshall-Lerner Condition Hold?” Journal of International Economics 31:301–316. . 2000. “One Money, One Market : the Effect of Common Currencies on Trade.” Economic Policy 30:7–47. Sapir, A., K. Sekkat, and A. Weber. 1994. “Tunisia-Export Growth: Determinants and Prospects.” Cepr discussion paper 1041, The Impact of Exchange-Rate Fluctuations on European Union Trade. Sauer, C., and A. Bohara. 2001. “Exchange Rate Volatility and Exports: Regional Differences between Developing and Industrialized Countries.” Review of International Economics 9(1):133–152. Sekkat, K. 1998. “Exchange Rate Variability and EU Trade.” Economic Papaers, vol. 127. Sekkat, Kh., and A. Vaoudakis. 1999. “Incentive Policies and Manufactured exports in North Africa.” ERF Working Paper, vol. 9922. Wang, K-Li., and C. Barrett. 2002. “A New Look At The Trade Volume Effects of Real Exchange Rate Risk.” Department of Applied Economics and Management, Working Paper, vol. 2002-41. Weliwita, A., E. Ekanayake, and H. Tsujii. 1999. “Real Exchange Rate Volatility and Sri Lanka’s Exports to the Developed Countries, 1978–96.” Journal of Economic Development 24, no. 1. Yotopoulos, P. 1996. “Exchange rate Parity for Trade and Development: Theory, Tests, and Case Studies.” Technical Report, Cambridge University Press. 33 Please note: You are most sincerely encouraged to participate in the open assessment of this discussion paper. You can do so by posting your comments. Please go to: http://www.economics-ejournal.org/economics/discussionpapers/2008-29 The Editor www.economics-ejournal.org © Author(s) 2008. This work is licensed under a Creative Commons License - Attribution-NonCommercial 2.0 Germany