PASS THROUGH OF EXTERNAL SHOCKS TO DOMESTIC PRICES IN TURKEY

Transcription

1 PASS THROUGH OF EXTERNAL SHOCKS TO DOMESTIC PRICES IN TURKEY Omer AKKUS, Anadolu University, Department of Economics Eskisehir, Turkey, oakkus@anadolu.edu.tr Abstract It is crucial to put forth the relation between external shocks and domestic prices because of the strong correlation between them in Turkey. This paper investigates pass through of external shocks to domestic prices in Turkey over the period In order to assess the impact of external shocks on price, we use McCarthy (1999) VAR model by employing monthly data set. According to the assumption of model, a variable in model can not be affected simultaneously by previous variable but it affects next variable. The empirical results indicate external shock is an important determinant of domestic prices according to results obtained from variance decomposition. A shock in import price affects industrial production index variance more significantly than oil price both in the short term and long term.. The pass through of a shock in import price index to wholesale price index is quicker than the pass through of oil prices to wholesale price index. Keyworlds : Pass Through, External Shocks, Domestic Prices, VAR models JEL Codes : C22, E31, F31 1. Introduction Exchange rate which is one of the most important variables for trade balance has a crucial role for macroeconomic stability. Therefore, it has always been a controversial issue in the literature. Exchange rate is an important part of transition mechanism not only because of traditional trade structure and input composition of production function but also because of balance of payment and the degree of financial system development (Başçı et. al, 2008). It is commonly believed that the pass through of external shocks to inflation is crucial to determine how external shocks affect domestic prices. The volume of pass through of external shocks to prices gives a clue to authoritiesm for the conduct of monetary policy (Devereux et. al, 2006). Central Banks have an important role on price stability.if the pass through of external shocks to domestic prices is low, central banks follow policy support instruments easier to maintain price stability. On the other hand, if the pass through of external shock to domestic prices is high, it may facilitate the spread of global crisis. Thereby, many central banks have hesitations while prefering the exchange rate regime to not involve in crisis (Calvo ve Reinhart, 2002). It has been reported in many studies that the pass through of exchange rate in based on law of one price. A product price without trade barriers is equal in both countries according to law of one price or purchasing power parity. Any changes in exchange rate provide product prices in domestic countries simultaneously. Pass through of exchange rate to prices generally refers to Dornbusch study (1989). Dornbusch had an analysis about fluctuations of domestic prices and exchange rate fluctuations by using industrial 1

2 organisation models. Dornbusch (1989) tried to find out how exchange rate fluctuations affect domestic prices and how imported goods affect the market power in country by taking into account market concentration ratio and substitution effect. it is very important to price out in terms of currency in the examining country or in terms of local currency in the country which produces goods when modeling pass through of exchange rate to prices and it plays a determinant role to choose the appropriate exchange rate regime (Choundhri et. al, 2002). McCarthy (1999) tried to explain pass through of exchange rate to prices by using chain models of distrubition. McCarthy emphasizes that an increase in exchange rate primarly affects imported goods after this it affects wholesale price index and in the last stage it affects consumer price index. In this study, McCarthy chain models of distrubition is going to be used the period for Turkey. Literature on the pass through of exchange rate is going to be in second part of study. Theoretical base and econometric model and method are going to be introduced in the third part. The last part of study consists of conclusion. 2. Literature Dornbusch (1987) reached the conclusion that the pass through of exchange rate is low when substitution of product in countries is high. According to Fisher (1989), market fragmentation is increased the pass through of exchange rate.. Menon (1995) suggests that the presence of multinational cooperation in the market reduces the pass through of exchange rate. Menon also suggests that the pass through of exchange rate is weaker under flexible exchange rate regime. Froot ve Klemperer (1989) researched whether the reflection of shocks to exchange rate is temporary or permanent by using two term model and tried to find out how the pass through of exchange rate become different. According to intertemporal model used by Froot and Klemperer, the pass through of exchange rate in second period depends on market share of companies in the initial period. Mann (1986) highlighted that the pass through of exchange rate is affected negatively by exchange rate volatility. Mann tried to indicate that the pass through of exchange rate fell under the high exchange rate volatility. McCarthy (1999) suggests that small economy has higher pass through of exchange rate than large economy. Ball (1999) and Svensson (1998) claimed that the openness of economy decreased the degree of the pass through of exchange rate to prices. Ho ve McCauley (2003) could not find any significant relationship between the pass through of exchange rate and openness of economy in their study. However, they put forth that lower degree of 2

3 income leads to higher degree of pass through of exchange rate. Devereux, Engel ve Storgaard (2003) reached the conclusion that countries which have more stable monetary policy and less exchange rate volatility are thought to price out in its currency by exporter and they suggest that the pass through of exchange rate to import price is lower. Coricelli, Jazbec ve Masten (2004) examined in their study some countries, Czech Republic, Poland, Hungary, Slovenia whether there is the pass through of exchange rate to prices or not by using VAR models. They found out that degree and speed of pass through of exchange rate to prices are high for Hungary and Slovenia while the pass through of exchange rate to prices is low for Czech Republic and Poland in the framework of exchange rate based stabilization policies and they concluded that exchange rate policy is an important determinant of pass through. Barhoumi (2006) analysed 24 developing countries by using panel cointegration method for the period annually and it is found different features of pass through within these developing countries. Holmes (2006) analysed 12 European Union Countries by using DOLS between-dimension approach with monthly data for the period and the results of estimation indicated that the pass through of exchange rate to prices in these countries decreased for the examined period. Mihaljek ve Klau (2001) examined the period for 13 developing countries and they found out that Turkey is one of the countries having most experienced pass through of exchange rate to prices among these developing countries. Arbatlı (2003) used VAR models in her study for the period between and the results indicated that the pass through of exchage rate to consumer price index decreased after Kara et. all (2005) observed changes of pass through coefficient of exchange rate over time and they found out that the pass through of exchange rate to consumer price index decreased approximately ten times over time. Aldemir (2006) put forth that the pass through of exchange rate to import prices was about 90 percent before 2001 but it was about 15 percent the period after Ito ve Sato (2007) investigated 8 developing countries and also Turkey included by using VAR models for the period montly data and they suggest that the pass through of exchange rate to consumer prices is 50 percent and it is quite high. Cazorzi et. all (2007) analysed 12 developing countries, United Stated, Japan and countries which use euro currency for the period quarterly data and it is found that the pass through of exchange rate to import prices is high but it is about 11 percent for Turkey and it is lower than other examined countries. 3

4 3. Method and Data McCarthy suggests that exchange rate, prices of imported goods, wholesale price index and consumer price index affects one another respectively. There is an assumption in the model and according to this assumption, a variable cannot be affected by a previous variable simultaneously but it affects the next variable which comes after itself. This study is based on McCarthy (1999) model. The model is constructed as the following ; (1) lop t = Ex t-1 [lop t ]+ ε t op (2) lindp t = Ex t-1 [lindp t ] +α 1 ε t op + εt indp (3) le t = Ex t-1 [le t ] +β 1 ε t op + β 2 ε t indp + ε t e (4) lipt = Ex t-1 [lip t ] +γ 1 ε t op + γ2 ε t indp +γ3 ε t e + εt ip (5) ltpi t = Ex t-1 [ltpi t ] +φ 1 ε t op + φ2 ε t indp +φ3 ε t e + φ4 ε t ip + εt tpi (6) lcpi t = Ex t-1 [lcpi t ] +φ 1 ε t op + φ2 ε t indp +φ3 ε t e + φ4 ε t ip + εt i,tpi + εt cpi Abbreviations refer to op ; oil prices, indp ; industrial production index, e ; exchange basket, ip ; import price index, tpi ; wholesale price index, cpi ; consumer price index. The study focuses on the period monthly data for Turkey by using VAR models. Lop; it is a proxy for logarithm of oil prices in Turkey. This proxy refers to the monthly average crude oil prices in United States multiplied by monthly average Turkish Lira/ Dollar. Lindo; logarithm of industrial production index (2010=100), it is seasonally adjusted by TRAMO/SEATs. le; logarithm of monthly average Turkish Lira/ Dollar +Turkish Lira/ Dollar. lip; logarithm of import price index (2005=100). ltpi ; logarithm of wholesale price index (2003=100). lcpi ; logarithm of consumer price index (2003=100). Before the estimation we have checked if the variables are stationary or not. Logarithm of level series are run for unit root tests and the series are I(1) according to Augmented Dickey-Fuller and Phillips-Perron tests. 4

5 Table 1. Variables ADF Phillips-Perron Oil Price * * Industrial production index * * Exchange Rate Import price index * * Wholesale price index * * Consumer price index * * * Sing indicates the rejection of unit root at 1% level It seems that the constructed models are a part of VAR model structure. Therefore, VAR models may reveal shocks by applying Cholesky decomposition. Cholesky decomposition makes use of requirements to identify VAR model (Hahn, 2003). By estimating VAR model and by applying Cholesky decomposition, impulse-response functions may be analysed through some indicators about inflation. Hereby, not only could degree of pass through be observed but also speed of it. Before estimating VAR model, lag lengths should be determined and the results of residual tests are also very important. Serial correlation test, heteroscedasticity test and normality test should be run on residual to indicate its diagnostics. Therefore, the pass through speed of external shocks could be analysed through impulse-response functions. Table 2. Lag LogL LR FPE AIC SC HQ NA * * * * * *Maximum lag lengths 5

6 Maximum lag lengths for VAR model are chosen according to Hannan-Quinn Information Criterion (HQ), Schwarz Information Criterion (SIC), Akaike Information Criterion (AIC) and Final Prediction Error (FPE). One lag of first differenced series is used in the model estimation. LM test confirms that there is not any autocorrelation problem in the constructed VAR model. Heteroscedasticity test indicates that variance of residuals are stationary in favour of assumption. As for the normality of residuals, it could not be rejected at 5% level and it is clear that the normality test results are satisfied. Impulse-Response Functions Figure 1. Accumulated Response of LCPI to Cholesky One S.D.LIP Innovation Figure 1 indicates that consumer price index positively respond to one percent increase in import prices and it is statistically significant. The pass through of import prices to consumer prices is about 21 percent after first six months and it seems stable after this month. Figure 2. Accumulated Response of LTPI to Cholesky One S.D. LIP Innovation Figure 2 displays that the pass through of import prices to wholesale price index is about 26 percent after first six months and it is stable after this six months. Figure 3. Accumulated Response of LINDP to Cholesky One S.D. LIP Innovation 6

7 Figure 3 indicates that industrial production index is positively affected by one percent change in import prices. The pass through of import price to industrial production index is about 36 percent after first eight months and it seems that the pass through of it is stable after this months. Figure 4. Accumulated Response of LINDP to Cholesky One S.D. LOILP Innovation Figure 4 indicates that the pass through of a percent increase in oil prices to industrial production index is about 7 percent after first eight months. It seems that the pass through of it is stable after this period. Figure 5. Accumulated Response of LTPI to Cholesky One S.D. LOILP Innovation Figure 6 shows that the pass through of a shock in oil price to wholesale price index is about 22 percent after seven months. The pass through of import price index to wholesale price index higher than to consumer prices. Impulse-response functions inform not only degree of pass through but also speed of pass through. The pass through of oil prices to industrial production index and wholesale price index occurs in nine months while approximately 90 percent of the pass through of import prices to consumer price index and wholesale price index occurs in six months. The pass through of a shock in import price index to wholesale price index is quicker than the pass through of oil prices to wholesale price index. Because import price index directly affects wholesale price index and consumer price index. 7

8 Variance Decomposition Although impulse-response functions contain some in formations about the speed and degree of pass through to domestic prices, it does not detect a shock occurring in domestic prices. Variance decomposition informs about importance of external shocks by decomposing according to endogenous variables in the VAR model including domestic prices of a shock occurring in the external variables. Figure 6. Variance Decomposition of Consumer Price Index Percent External shock is a crucial determinant of variance decomposition of consumer price index. External shocks consisting of 13 percent of import prices and 9 percent of oil prices explain approximately 22 percent of consumer prices in total. Figure 6. Variance Decomposition of Industrial Production Index Percent A shock in import price affects industrial production index variance more significantly than oil price both in the short term and long term. Approximately 27 percent of industrial production index variance is explained by external shocks. 8

9 Figure 6. Variance Decomposition of Wholesale Price Index Percent While approximately 7% of wholesale price index variance is explained by shocks in oil price, 16% of external shocks is explained by import prices. Briefly stated, external factors and shocks explain a very large proportion of the wholesale price index, consumer price index and industrial production index variance. Conclusion In order to estimate VAR model and to apply Cholesky decomposition enable to identify externa shocks. Therefore, impuse-response of some indicators about domestic to shocks prices have been investigated in the study.informations about the speed and degree of pass through of external shocks to domestic prices are obtained from impulse-response functions. Impulse-response functions inform not only degree of pass through but also speed of pass through. The pass through of a shock in import price index to wholesale price index is quicker than the pass through of oil prices to wholesale price index. Variance decomposition informs about importance of external shocks by decomposing according to endogenous variables in the VAR model including domestic prices of a shock occurring in the external variables. External shock is an important determinant of domestic prices according to results obtained from variance decomposition. A shock in import price affects industrial production index variance more significantly than oil price both in the short term and long term. 9

10 REFERENCES Arbatlı, Elif C. (2003), Exchange Rate Pass Through in Turkey: Looking for Asymmetries, Central Bank Review (2), Aldemir, Şenkan (2006), Türkiye Ekonomisinde Döviz Kurunun Yurtiçi İthalat Fiyatlarına Geçiş Etkisi: , Uluslararası Ekonomi ve Dış Ticaret Politikaları, 1(2), Barhoumi, K. (2006), Differences in Long Run Exchange Rate Pass-Through into Import Prices in Developing Countries: An Empirical Investigation, Economic Modeling, 23 (6), Ball, Laurence M. (1999), Policy Rules for Open Economies, In: Taylor, John B. (Ed.), Monetary Policy Rules. Chicago: University of Chicago Press Bascı, E., Özel, Ö. ve Sarıkaya Ç. (2008), The Monetary Transmission Mechanism in Turkey: New Developments in Transmission Mechanisms for Monetary Policy in Emerging Market Economies, BIS Papers No: 35, Calvo, Guillermo A. ve Carmen M. Reinhart (2002), Fear of Floating. Quarterly Journal of Economics, 117(2), CaZorzi Michele, Elke Hahn ve Marcelo Sanchez (2007), Exchange Rate Pass-through In Emerging Markets. European Central Bank Working Paper No: 739 Coricelli, F., Jazbec, B. ve Masten, I. (2004), Exchange Rate Policy and Inflation in Acceding Countries: The Role of Pass-through, William Davidson Institute Working Paper No: 674 Choundhri Eshan U., Faruqee Hamid ve Hakura Dalia S. (2002), Explaining the Exchange Rate Pass-Through in Diffrent Prices, IMF Working Paper, WP/02/224, 1-33 Devereux, M., Engel C. ve Storgaard P. (2003), Endogenous Exchange Rate Passthrough When Nominal Prices are Set in Advance. NBER Working Papers No: 9543 Devereux, M., Lane P. ve Xu J. (2006), Exchange Rates and Monetary Policy in Emerging Market Economies, The Economic Journal, Volume 116, Issue 511, Dornbusch, Rudiger (1987), Exchange Rates and Prices, American Economic Review 77, Dornbush, Rudiger (1989), Exchange Rates and Inflation, London, The MIT Press. Hahn, Elke (2003), Pass-Through of External Shocks to Euro Area Inflation, Working Paper.243 Holmes, M. (2006), Is A Low-Inflation Environment Associated With Reduced Exchange Rate Pass Through? Finnish Economic Papers, 19 (2), Froot, Kenneth A. ve Paul Klemperer. (1989), Exchange Rate Pass-Through When Market Share Matters, NBER Working Paper No: 2542 Ho, Corrinne ve Robert N. McCauley. (2003), Living with Flexible Exchange Rates: Issues and Recent Experience in Inflation Targeting Emerging Market Economies, BIS Working Paper No.130 Ito, Takatoshi ve Kiyotaka Sato (2007), Exchange Rate Pass-Through and Domestic Inflation: A Comparison between East Asia and Latin American Countries Mann, Catherine L. (1986), Prices, Profits Margins and Exchange Rates, Federal Reserve Bulletin 72, McCarthy, J (1999), Pass Through of Exchange Rates and Import Prices to Domestic Inflation in Some Industrialized Economies, BIS Working Paper No: 79,