Charles University in Prague Faculty of Social Sciences Institute of Economic Studies. University of Strasbourg Institute of Political Studies

Transcription

1 Charles University in Prague Faculty of Social Sciences Institute of Economic Studies University of Strasbourg Institute of Political Studies MASTER'S THESIS The Determinants of Inflation Differentials across Central and Eastern European Countries Author: Mihaela Gurbulea Supervisor: doc. Roman Horváth, Ph.D. Academic Year: 2014/2015

2 ii Declaration of Authorship The author hereby declares that he compiled this thesis independently, using only the listed resources and literature, and the thesis has not been used to obtain a different or the same degree. The author grants to Charles University permission to reproduce and to distribute copies of this thesis document in whole or in part. Prague, July 31, 2015 Signature

3 iii Acknowledgments It is with immense gratitude that I would like to thank my supervisor, doc. Roman Horváth, Ph.D., for his useful advice and invaluable guidance throughout the process of writing this thesis. I would also like to express my profound gratitude to my parents for their love and unconditional support.

4 iv Abstract The thesis aims at identifying the reasons behind the heterogeneous inflation performance of countries across Central and Eastern Europe. The impact of a large number of variables is being assessed in a dynamic panel data model covering 20 countries over the period The empirical results suggest that cross country differences in inflation are attributed to the structure of the economy, to the capital deepening effects and openness. Along with the structural factors, cyclical positions also prove to be of particular importance in explaining inflation across the region, since during the last decade most of the Central and Eastern European countries have experienced fast GDP growth, a credit boom and increased domestic demand that in turn fueled inflation. JEL Classification Keywords E31, E62, P24 Inflation differentials; Central and Eastern Europe; dynamic panel data model; biascorrected LSDV. Author s e mail Supervisor s e mail mihaela.gurbulea@gmail.com roman.horvath@fsv.cuni.cz

5 v Contents List of Tables... vii List of Figures... viii Acronyms... ix Master's Thesis Proposal... xii 1 Introduction Inflation Dynamics in Central and Eastern European Countries Evolution of Inflation Inflation differentials across CEE countries Theories and Factors of Inflation Structural factors of inflation differentials Price convergence Tradable goods price convergence Non tradable goods price convergence External factors Business cycle factors Data and Methodology Data Description Empirical Methodology Empirical Results Separate regressions with groups of factors Estimation results for the structural variables Estimation results for the business cycle variables Results for the institutions and policies variables Estimation results for the economic shocks From General to Specific... 56

6 vi 6 Conclusion Bibliography Appendix A: Inflation rates in the euro area and CEEC Appendix B: Data Sources Appendix C: Summary Statistics... 69

7 vii List of Tables Table 2.1: Pre EU and Post EU summary inflation statistics for the NMS ( ) Table 2.2: Summary inflation statistics for transition and candidate countries ( ) Table 2.3: Inflation differentials in CEEC relative to the euro area Table 3.1: Correlation between inflation and openness in nine Eastern European countries ( ) Table 3.2: Exchange rate regimes in SEE countries Table 4.1: Panel unit root tests results for the inflation rate Table 5.1: The impact of structural factors on the inflation in CEEC (LSDVC estimation results) Table 5.2: Correlation matrix for the EBRD transition indicators Table 5.3: The impact of transition indicators on the inflation in CEEC (LSDVC estimation results) Table 5.4: The impact of business cycle factors on the inflation in CEEC (LSDVC estimation results) Table 5.5: The impact of institutions and policies on inflation in CEEC (LSDVC estimation results) Table 5.6: The impact of economic shocks on the inflation in CEEC (LSDVC estimation results) Table 5.7: The impact of selected variables on the inflation in CEEC (LSDVC estimation results) Table A.1: Inflation in the euro area and CEEC over Table B.1: Variables, their definitions and the corresponding data sources Table C.1: Summary statistics for the included variables... 69

8 viii List of Figures Figure 2.1: Average inflation rates in CEEC and Euro area Figure 2.2: Inflation in CEE per groups of countries Figure 2.3: Average inflation rates in CEEC over Figure 2.4: Inflation differentials in CEEC relative to the euro area (in percentage points) Figure 2.5: Inflation differentials in Central, Baltic and Southeastern countries ( ) Figure 2.6: Inflation differentials in CIS West and candidate countries ( ) Figure 2.7: Cumulated inflation differentials in CEE countries Figure 3.1. Unit labor costs in manufacturing (annual growth in percent) Figure 5.1: Inflation vs. Imports in percent of GDP in CEE countries Figure 5.2: Inflation vs. GDP growth in CEE countries ( )

9 ix Acronyms Country Abbreviations: AL BA BG BY CZ EE HR HU LT LV MD ME MK PL RO RS RU SI SK UA Albania Bosnia and Herzegovina Bulgaria Belarus Czech Republic Estonia Croatia Hungary Lithuania Latvia Moldova Montenegro FYR of Macedonia Poland Romania Serbia Russia Slovenia Slovakia Ukraine Other Abbreviations: AB BB AR BS CEE CEEC CIS Arellano Bond estimator Blundell Bond estimator Arellano Bond test for auto correlation in the residuals Balassa Samuelson Central and Eastern Europe Central and Eastern European Countries Commonwealth of Independent States

10 x CPI EA EBRD ECB EMU ERM II EU GDP GMM ID ILO IMF IT NMS LSDV LSDVC OMS SEE USD WDI WEO WGI Consumer Price Index Euro Area European Bank for Reconstruction and Development European Central Bank Economic and Monetary Union Exchange Rate Mechanism II European Union Gross Domestic Product General Method of Moments Inflation Differentials International Labour Organisation International Monetary Fund Inflation Targeting New Member States Least Squares Dummy Variable Bias corrected Least Squares Dummy Variable Old Member States Southeastern Europe United States Dollar World Development Indicators World Economic Outlook Worldwide Governance Indicators

11 Master's Thesis Proposal Author: Mihaela Gurbulea Supervisor: Doc. Roman Horvath Ph.D. E mail: mihaela.gurbulea@gmail.com E mail: roman.horvath@fsv.cuni.cz Phone: Phone: Specialization: Corporate Finance and Strategy Defense Planned: June 2015 Proposed Topic: The determinants of inflation differentials across Central and Eastern European Countries. Motivation: In the last decade there has been a significant gap in the inflation performance across CEE countries, namely between the New Member States (NMS) of the European Union and the developing economies in Eastern Europe. NMS managed to successfully complete their transition process and attained low levels of inflation required by the Maastricht Treaty. On contrary, Western CIS countries such as Moldova, Belarus, Russia and Ukraine are still dealing with high inflation. In my thesis I aim at finding the factors underlying the divergent paths of inflation in transition economies compared to NMS. Understanding what drives inflation in these economies is important for several reasons. First, high inflation reduces welfare. Second, transition economies are exposed to a larger number of shocks, both endogenous and exogenous. In case of an external shock, different policies may be implemented depending on its degree of persistence. Third, high inflation reduces international competitiveness and may have adverse implications on output. Hammermann & Flanagan (2007) explain the inflation differentials in transition economies by institutional factors. Their results suggest that the motivation for central banks to target a higher inflation in CIS countries is due to differences in the degree of price liberalization, openness and fiscal environment. The importance of institutional factors in driving inflation in transition countries has been also supported by Cottarelli, Griffiths & Moghadam (1998). Other factors with significant effect on inflation proved to be the fiscal deficits, relative price changes and the exchange rate regime. However, these studies are based on old data and different sample of transition countries. Hypotheses: 1. Hypothesis #1: Lower degree of liberalization and structural reforms increase central banks incentives to inflate in transition countries. 2. Hypothesis #2: Transition economies are more exposed to exchange rate risks and thus, have higher inflation. 3. Hypothesis #3: Fiscal environment accounts for differentials in inflation among countries. Methodology: I will collect the data from the International Monetary Fund database, because it is available for all countries and this will help to ensure comparability across countries analyzed in the study. Also, I plan to work with EBRD transition indicators to capture the effect of liberalization and structural reforms on inflation differentials. For modelling the inflation

12 xiii differential, I will follow the approach of Hammermann & Flanagan (2007) and I will use panel data econometric techniques in order to exploit both cross country and time dimensions of the data. Hammermann & Flanagan (2007) estimate determinants of inflation using time fixed effects. For adressing the sources of inflation differentials I will use a decomposition to capture the contribution of each variable in percents. Expected Contribution: I plan to contribute to the existing literature in several aspects. First, I will use latest data in order to focus on the more recent inflation performance of the countries analyzed in the study. Unlike older studies that focus on different sample of transition countries that by now have completed the transition process, I will update the country classifications. Second, I will account for the influence of a bigger range of variables on the inflation differentials across CEE region. Third, I m interested in quantifying the impact of the global financial crisis on the inflation performance across the analyzed countries and see if they reacted differently or not to this shock. Outline: 1. Introduction motivate the importance of analysing the inflation differentials in CEE countries. 2. Inflation dynamics in CEEC I will analyze the patterns and characteristics of inflation and inflation differentials across the economies in the region. 3. Theories and factors of inflation will discuss theories explaining inflation and make a survey of the empirical literature investigating its drivers. 4. Data and Methodology I will model the inflation differential and motivate the use of the variables. 5. Results present the results of the estimation and compare them with other findings in the literature. 6. Conclusion provide concluding remarks about the findings and their policy implications. Core Bibliography: Cottarelli, C., Griffiths, M. & Moghadam, R. (1998): The Nonmonetary Determinants of Inflation: A Panel Data Study. International Monetary Fund Working Papers, no. 98/23. Cukierman, A., Miller, G.P. & Neyapti, B. (2002): Central Bank Reform, Liberalization and Inflation in Transition Economies an International Perspective. Journal of Monetary Economics 49: pp Hammermann, F. & Flanagan, M. (2007): What Explains Persistent Inflation Differentials across Transition Economies? International Monetary Fund Working Paper, no. 189/07. Horvath, R. & Koprnicka, K. (2008): Inflation Differentials in New EU Member States: Empirical Evidence. Czech Journal of Economics and Finance 58 (7 8): pp Mihaljek, D. & Klau, M. (2008): Catching up and Inflation in Transition Economies: the Balassa Samuelson Effect Revisited. BIS Working Paper, no Staehr, K. (2009): Inflation in the New EU Countries from Central and Eastern Europe: Theories and Panel Data Estimations. European Economy Occasional Papers 50: pp Author Supervisor

13 Introduction 14 1 Introduction In the aftermath of the dissolution of the Soviet Union and the breakdown of the command economies, the countries from Central and Eastern Europe have been subject to high levels of inflation. The period of transition to market economy was characterized by the implementation of structural reforms such as price and trade liberalization, and currency devaluations that have exerted upward pressure on inflation. Still, the New EU Member States managed to successfully complete their transition process and attained low levels of inflation, while Western CIS countries such as Moldova, Belarus, Russia and Ukraine are still dealing with high and volatile inflation. As a result, the successful approach of the NMS to marketisation has been reflected into their accession to the EU. Meanwhile, the Balkan countries are still standing to join the EU and have been granted the candidate status, while Moldova and Ukraine signed the EU Association Agreement in 2014 and are actively pursuing EU membership. Therefore, understanding the reasons behind this significant gap in the inflation performance of these countries is particularly important for their future prospects, since the intention to join the EU will bring the responsibility of achieving price stability. In the thesis, we aim at finding the factors underlying the divergent paths of inflation across countries from CEE. Understanding what drives inflation in these economies is important for several reasons. First, high inflation reduces welfare. Second, developing economies are exposed to shocks that may create instability and translate into volatile inflation. Third, knowing the drivers of inflation will help authorities in designing appropriate policies to ensure price stability. The literature studying inflation determinants across CEE countries is limited, but growing: Hammermann & Flanagan (2007) explain the inflation differentials in transition economies by institutional factors. Their results suggest that the motivation for central banks to target a higher inflation in CIS countries is due to differences in the degree of price liberalization, openness and fiscal environment. Horvath & Koprnicka (2008) analyze the contribution of the exchange rate, price levels and cyclical factors in explaining the inflation differentials in the New EU Member States vis-a-vis the euro area. Their estimation results point out that the highest contribution in explaining the inflation differentials across these countries is attributed to the factors related to the real convergence process. The factors influencing inflation in

14 Introduction 15 the NMS from Central and Eastern Europe are also investigated by Staehr (2009). However, all the afore mentioned studies focus on the inflation determinants only of the New Member States of the EU, whereas the research on the CIS West and the candidate countries inflation performance is very scarce. I contribute to the existing literature in several aspects. First, unlike older studies that focus only on the inflation in the NMS from CEE, I widen the country coverage by adding the Balkan countries and the CIS West countries. Second, I use the latest data in order to focus on the more recent inflation performance of the countries analyzed in the study. Third, I account for the impact of a bigger range of variables on the inflation differentials across the CEE region. The objective of our empirical exercise is to identify the main determinants of inflation in CEE countries and to investigate the reasons for which CIS West countries have the highest inflation across the region. The dataset is composed of annual data on 20 countries from Central and Eastern Europe covering the period As we have a wide range of variables that due to multicollinearity issues cannot be introduced in the regression in the same time, we opt for the general to specific modeling that consists of separately testing the influence of all variables and then eliminating those that do not have sufficient explanatory power. We estimate the models using the bias corrected Least Squares Dummy Variable estimator for unbalanced panels developed by Bruno (2005). Following this introduction, the rest of the thesis is organized as follows: Section 2 shows the evolution and characteristics of inflation and inflation differentials across the CEE countries; Section 3 discusses the theoretical framework about the determinants of inflation and reviews the empirical literature on this topic; Section 4 describes the data and the methodology used and Section 5 discusses the results. Finally, Section 6 concludes.

15 Inflation Dynamics in Central and Eastern European Countries 16 2 Inflation Dynamics in Central and Eastern European Countries 2.1 Evolution of Inflation Over the last decade, a sizeable gap has developed in the inflation performance of countries across Central and Eastern Europe. While there are significant differences between economies in this region and the evolution of inflation has country specific features, the New Member States of the European Union have converged to a lower level of inflation compared with the transition economies of CIS West (Belarus, Moldova, Russia and Ukraine) (Figure 2.1). Therefore, before proceeding to the investigation of possible reasons underlying this divergence, we will start this section by giving the big picture regarding the trends of inflation over the last decade in NMS, candidate countries and the transition economies of Western CIS Euro area EU-Central EU-Baltic CIS-West EU-Southeast Candidates Figure 2.1: Average inflation rates in CEEC and Euro area. Source: IMF World Economic Outlook In order to better assess the patterns of inflation in the region, we decided to divide CEE countries into five groups following the approach of Hammerman & Flanagan (2007). Figure 2.2 depicts the inflation performance of 20 CEE countries over the period In all NMS, except Slovenia, Slovakia and Romania, inflation increased in 2004 after the EU accession, but later it declined. On contrary, the inflation across the three Baltic States has continuously risen until Prior to

16 Inflation Dynamics in Central and Eastern European Countries 17 the global financial crisis there has been a significant increase in inflation across the whole region, which was due to the credit boom that gained weight in Once the crisis got underway, the average inflation in CEE has dropped to its lowest level since transition (Nath & Tochkov, 2010) EU-Central Czech Republic Hungary Poland Slovak Republic Slovenia 16 EU-Baltic Estonia Latvia Lithuania EU-Southeast Bulgaria Croatia Romania

17 Inflation Dynamics in Central and Eastern European Countries Candidates Albania Bosnia and Herzegovina FYR Macedonia Montenegro Serbia CIS-West Belarus Moldova Russia Ukraine Figure 2.2: Inflation in CEE per groups of countries. Source: IMF World Economic Outlook The inflation profile of candidate countries proved to be satisfactory, and moreover, the inflation rates were close to the euro area average (Figure 2.1). A clear exception in this group of countries is Serbia, whose inflation rate was fluctuating from double digit to single digit values (Table A.1). 1 Notably, the most stable inflation was observed in Albania. On contrary, the Eastern transition economies proved to be less successful in dealing with high prices and did not manage to stabilize them, which led to inflation levels on average above 10%. From the Figure 2.2 we can notice how in 2011 the inflation in Belarus started to diverge and attained a skyrocketing level of 53.23% in 2012 (Table A.1). As Figure 2.3 shows, the highest average inflation rates in CEEC during the last decade were observed in Belarus (19.82%) along with Ukraine, Russia and 1 See Appendix A.

18 % Inflation Dynamics in Central and Eastern European Countries 19 Moldova, that attained levels of 10.11%, 9.48% and respectively, 8.55%. Macedonia and Czech Republic reached the lowest inflation rates over the period (2.54% and 2.55%) BY UA RS RU MD RO LV BG HU EE LT ME SK PL HR SI AL BA CZ MK Figure 2.3: Average inflation rates in CEEC over Note: CPI (2010 = 100) was used instead of averaging the inflation rates. If CPI grew at this annual rate, compounded from 2003 to 2013 it would result in the same inflation rate that occurred in Source: Author s calculations based on World Bank data. The heterogeneous inflation performance across CEEC in the last decade suggests that inflation has been a key issue of transition for Eastern economies. Nonetheless, the Central European countries have also been subject to a transition process after the collapse of the communist regime in 1989, but thanks to the adoption of different policy approaches, they proved to be more successful in completing the transition. For example, Czech Republic, Hungary and Poland managed to restructure fast their economies, while the Southeastern Europe and former CIS implemented the economic reforms slower and less efficiently (Ackrill & Coleman, 2012). As a result, their successful approach to marketisation has been reflected into the accession to the EU. This fact indicates that the inflation dynamics in these economies is very important, as the commitment to work towards EMU membership has brought the responsibility of attaining low levels of inflation laid down by the Maastricht criteria. Indeed, we notice that inflation in NMS has significantly decreased compared to pre EU levels (Table 2.1). In the same time, only four countries have been able to fulfill the convergence criteria and adopt the euro (Slovenia, Slovakia, Estonia and Latvia). This fact certainly emphasizes the immensity of the challenges these countries have been facing.

19 Inflation Dynamics in Central and Eastern European Countries 20 Table 2.1: Pre EU and Post EU summary inflation statistics for the NMS ( ). Country Mean Std. Deviation Minimum Maximum Pre Post Pre Post Pre Post Pre Post Czech Republic Hungary Poland Slovakia Slovenia Estonia Latvia Lithuania Bulgaria Romania Source: Author s computations based on IMF data. In order to face the challenge of controlling inflation, one common approach in the monetary policies of CEEC was the adoption of inflation targeting. In Czech Republic, following the turbulences of the exchange rate and the eventual abandoning of the peg in 1997, the National Bank decided to pursue the inflation targeting. Czech Republic was the first transition economy to adopt the IT regime and it was implemented only after the other monetary policy regimes failed (Šmídková, 2008). In 1998 and 2001, the monetary authorities in Poland and Hungary started to lay the groundwork for the IT framework. This came as a consequence of unsuccessful efforts to use nominal anchors for managing inflation. In Slovenia, the monetary policy was focused on targeting the money supply in terms of M1 and by 2007 it became the first CEEC to join EMU without explicitly targeting inflation. The second member of CEEC that entered euro area was Slovakia in Estonia introduced its currency board in 1992 and after joining EU it has hard pegged its currency the kroon against the euro central rate. Initially, Estonia was planning to become an EMU member in 2008, but due to a high inflation experienced that year, it adopted the euro not earlier than by January Latvia, after joining ERM II in 2005, also adopted an exchange rate targeting by allowing its currency to fluctuate within 1% margin versus euro. As its inflation rates were missing the EMU benchmark, Latvia could join the euro zone only in Lithuania is another Baltic

20 Inflation Dynamics in Central and Eastern European Countries 21 state that had established its currency board arrangement but eventually it resulted into failure to comply with the euro adoption criteria. After experiencing a severe banking crisis in 1996, Bulgaria began to strictly maintain its currency against the Deutsche mark and later against euro and, in this way, it managed to reduce drastically its inflation rate. During inflation averaged 101%; however, this figure was biased upwards due to skyrocketing inflation in Likewise, Romania s inflation rate exhibited a considerable improvement taking into account the hyperinflation experienced in 1997 and the severe financial crisis from Its average inflation rate dropped from the level of 44.1% prior to EU integration to 5.36% during Table 2.2: Summary inflation statistics for transition and candidate countries ( ). Country Mean Std. Deviation Minimum Maximum Belarus Moldova Russia Ukraine Albania Bosnia and Herzegovina Croatia Macedonia Montenegro Serbia Source: Author s computations based on IMF data. Table 2.2 provides an overview of the inflation distribution during the last decade in the CIS transition countries and the potential candidates to the EU. When comparing the values of the Western CIS economies with those of NMS from Table 2.1, it is hard to overlook the discrepancies between these groups of countries. Precisely, the magnitude of inflation in the CIS group is emphasized by high average inflation rates that are above 8%. The maximum values of inflation across the region were observed in Belarus (59.2%) and Ukraine (25.2%). Russia and Moldova experienced an average inflation around 9%; however, the standard deviation is lower

21 Inflation Dynamics in Central and Eastern European Countries 22 compared to the ones of Belarus and Ukraine. Meantime, the candidates and potential candidates proved to be more successful in managing inflation and this fact is supported by the low average levels per country that did not exceed 4%, except Serbia, whose average inflation rate was 9.12%. 2.2 Inflation differentials across CEE countries Understanding the size of inflation differentials in each of the CEE countries is crucial for the further investigation of their determinants. In this section we are interested in comparing the inflation performance of NMS with the transition and candidate countries relative to the euro area. Figure 2.4 depicts the average deviation of CEE countries vis à vis the euro area over the period BY RU UA RS MD RO LV HU BG EE ME SK LT SI HR AL PL BA MK CZ Figure 2.4: Inflation differentials in CEEC relative to the euro area (in percentage points). Note: Countries are arranged in descending order according to the average differential from euro area over Source: Author s computations based on IMF data. In order to better distinguish the details concerning the distribution of inflation differentials, we will separate the countries into groups. As it can be seen from the Figure 2.5, the lowest average deviations relative to the euro area were observed in the EU Central and Baltic countries. Nevertheless, the extreme values of deviations were larger in the case of Baltic countries, mainly concerning the maximum values, i.e. Latvia had a maximum deviation of almost 12 percentage points relative to the euro area. The same thing can be pointed out about the Southeastern countries that had slightly higher average differentials vis à vis the euro area, but a wider amplitude; due to high levels of inflation experienced in the

22 Inflation Dynamics in Central and Eastern European Countries 23 late 90s and early 00s, Romania managed to reduce its inflation to single digit values only after EU-Central HU SK SI PL CZ EU-Baltic LV EE LT EU-Southeast RO BG HR Figure 2.5: Inflation differentials in Central, Baltic and Southeastern countries ( ). Source: Author s computations based on IMF data.

23 Inflation Dynamics in Central and Eastern European Countries 24 Russia, Moldova and Ukraine reached inflation differentials on average below 8 percentage points, whereas the inflation differentials of Belarus attained the mean value of percentage points and the maximum value of percentage points (Figure 2.6). As we saw in the previous section, the inflation path of candidates and potential candidates was close to the euro area average. The inflation in Montenegro, Albania, Bosnia and Herzegovina and Macedonia did not deviate from the euro area average by more than 2 percentage points. From the Figure 2.6 we can also notice how spread are the inflation differentials of Serbia, which extreme values lie within the minimum of 0.81 percentage points and the maximum of percentage points CIS-West BY RU UA MD Candidate countries RS ME AL BA MK Figure 2.6: Inflation differentials in CIS West and candidate countries ( ). Source: Author s computations based on IMF data. Figure 2.7 reports the cumulated inflation differentials relative to the euro area average for 19 CEE countries. For providing a more explicit view of the picture we decided to omit Belarus due to its large cumulated inflation differential that distorts the scale.

24 Inflation Dynamics in Central and Eastern European Countries High cumulated ID Moderate cum. ID Low cum. ID Figure 2.7: Cumulated inflation differentials in CEE countries. Source: Author s computations based on IMF data. As it can be seen from the Figure 2.7 and Table 2.3, the CEE countries can be divided into three groups according to the patterns and size of their cumulated inflation differentials. The first group includes the countries which inflation differentials vis à vis the euro area were on average lower than 1 percentage point and have not experienced deviations higher than this value for longer than two consecutive years. This group of economies includes the Czech Republic, Slovenia, Poland, Croatia and the candidates Albania, Bosnia and Herzegovina and Macedonia. Table 2.3: Inflation differentials in CEEC relative to the euro area CZ HU PL SK SI EE LV LT BY MD RU UA AL BA BG HR MK ME RO RS Source: Author s computations based on IMF data.

25 Inflation Dynamics in Central and Eastern European Countries 26 Regarding the countries with relatively moderate cumulated inflation differentials Hungary, Estonia, Latvia, Bulgaria and, to a lesser extent, Slovakia, Lithuania and Montenegro the inflation differentials have exceeded on average 1 percentage point. Nonetheless, there are some notable differences between the above mentioned countries. Slovakia, Lithuania and Montenegro have experienced average levels of inflation differentials slightly above 1 percentage point. In contrast, the inflation differentials in Hungary have been persistently positive and high since 2003 and recorded an average level of 2.81 percentage points. Similarly, the Bulgarian differentials have stayed above the euro area figures except in Latvia s inflation has risen sharply until 2008, recording an percentage points differential and declining afterwards to levels below the euro area average. Estonia maintained positive differentials with regard to the euro area. Yet, the negative inflation differentials experienced in 2003 and 2009 did not offset the high positive differentials across the period, leading to an average deviation of 2.14 percentage points. Finally, the third group faced high and volatile inflation differentials over the whole decade. Among the economies that suffered from high cumulated inflation differentials are Romania, Serbia and the Western CIS countries. In Belarus, the inflation was persistently high and attained its maximum differential from euro area in 2012 (56.72 percentage points). All the countries in this group experienced over the entire period large positive inflation differentials with few exceptions Moldova in 2009 and Ukraine in had inflation levels below the euro area figure. Romania succeeded in reducing its inflation differentials by 2007; however, it started to increase again after the EU accession. The inflation in Serbia was volatile vis à vis the euro area and the differentials showed repeatedly episodes of increase followed by declines. Overall, its inflation differentials averaged 7.07 percentage points. Summing up, there has been a substantial and persistent inflation gap between countries across Central and Eastern Europe. There is a noticeable difference in the size of the cumulated inflation differentials among economies and this is not surprising given that variations in cross country inflation are common even in a monetary union due to differences in business cycles and productivity growth (Praet, 2012). In spite of that, the inflation differentials become a matter of concern when they are highly persistent and when the differences go beyond the mismatch of business cycles.

26 Theories and Factors of Inflation 27 3 Theories and Factors of Inflation The previous section gave the big picture of the evolution and persistence of inflation across countries in Central and Eastern Europe; however, this is not sufficient to understand the particular importance of the inflation differentials for the development of economic policies. For this purpose, we will proceed with analyzing the underlying reasons for inflation differentials along with reviewing the literature investigating them. Following the approach of ECB (2012), we will distinguish the factors according to the timing of their effects: long term or structural factors, medium term or factors related to the business cycle and short term or one off factors. 3.1 Structural factors of inflation differentials Price convergence One of the structural factors that contribute to the differences in inflation across countries is the price level convergence. It implies that if cross country prices expressed in common currency are initially different, countries where prices were initially low will be subject to a higher inflation such that, eventually, prices across countries will converge. The literature distinguishes between the price convergence of tradable and non tradable goods Tradable goods price convergence As far as the convergence of tradable goods is concerned, the literature explains it through the purchasing power parity theory. It assumes that the prices of a representative basket of products expressed in common currency should not differ across countries. Therefore, the establishment of the European Single Market in 1992 may give rise to the expectation that prices will converge across the member states, at least for traded goods. Moreover, the Eastern enlargement of EU is of great importance for the completion of the market integration and has received much attention from both the academics and policy makers. The question is whether the lower initial prices in NMS and their ongoing catch up process would cause an increase in the price dispersion within EU (Egert, 2007).

27 Theories and Factors of Inflation 28 Dreger et al (2007) investigate the effects of the EU 10 enlargement on the price convergence in the European market and find the presence of β convergence 2. Their empirical results suggest that the price differentials across countries are removed by only 6% each period and the speed of convergence increases with the tradability of the goods. According to their study, the speed of β convergence proves to be higher in the case of basic headings given that they include more homogenous products. Additionally, they observe a reduction in the price dispersion over time (σ convergence), but the speed of σ convergence declines after the EU enlargement in In view of the σ convergence, Lindenblatt & Feuerstein (2014) point out that the Eastern enlargement had a positive effect on reducing the deviation of food prices across Old and New Member States and this was mainly attributable to the convergence between these two groups of countries, rather than to the convergence within them. On the contrary, Wolszczak Derlacz & De Blander (2009) reject the presence of σ convergence within the OMS and three NMS (Czech Republic, Poland and Hungary) and do not find any decline in their price differentials in Given that their data are only until 2006, their results thus cannot provide further insights into the impact of the EU enlargement on the price convergence in the NMS. Solakoglu & Civan (2006) examine the wheat price convergence of transitional economies to the world average prices. They identify that the EU membership prospect played a crucial role in the fast convergence of CEE countries compared to the CIS and SEE countries. Solakoglu & Civan (2006) conclude that it was the guidance received from EU along with the large investment inflows that enabled CEE to successfully integrate the world markets. Besides investigating whether the inflation differentials across countries are due to the price level convergence, it is also important to distinguish how the convergence works. If it occurs mostly through tradable goods, the divergence between countries may be transitory; but if the price dispersion is owed to the differences in productivity and living standards, the convergence can last longer (Rogers, 2001). 2 β convergence implies that countries with initially low prices will experience higher inflation, converging eventually to the mean (Barro & Salla i Martin, 1992).

28 Theories and Factors of Inflation Non tradable goods price convergence The progress towards an integrated market in Europe through the four freedoms (free movement of goods, services, capital and labor) may decrease price differences across countries not only for traded goods, but also for the non-tradable goods. This price convergence can be explained by the Balassa Samuelson effect. 3 This hypothesis explains the price levels divergence through the differences in productivity growth between the tradable and non tradable goods sectors. It implies that whenever there is a productivity growth in the traded goods sector, relative wages rise accordingly. As low price countries increase faster their productivity in the traded goods than in the non traded goods sector, this puts upward pressure on wages without increasing the price of the goods produced. Given the labor mobility across sectors, the wages in the non tradables sector will rise too. In turn, this will push up prices in the countries with initially low price levels, leading to higher inflation. In hopes that the Balassa Samuelson hypothesis provides a plausible reason why fast growing economies experience higher level of inflation, many economists sought to investigate this effect. Dreger et al (2007) obtain high coefficients for the catching up regressor captured through the income per capita, productivity and wages. Its effect is particularly important for NMS in the non traded goods sector, which is in line with the BS model. In the attempt to investigate the drivers of inflation differentials across CEE countries by accounting for a wide range of variables, Staehr (2009) concludes that the differences in labor productivity in the traded and non traded goods sectors contributed to the price convergence within these countries, driving inflation up. Moreover, the BS effect proved to be robust across different subsamples. On the contrary, Egert (2007) finds no evidence of the BS effect on the inflation rates of transition countries and indicates that the price convergence between euro area countries and CEEC may be influenced by other structural factors such as the shift towards higher quality products or the role of the distribution sector. Similarly, the BS effect turned to be insignificant in explaining the divergent paths of CEE and CIS countries in a study performed by Hammerman & Flanagan (2007). The heterogeneity of the empirical results on the BS effect on price convergence suggests that it does not fully explain the differences in productivity 3 See Balassa (1964) and Samuelson (1964).

29 Theories and Factors of Inflation 30 growth across sectors and countries. If the productivity is equal in the manufacturing and service sector, convergence can occur without the BS effect (ECB, 2003) External factors The divergent paths of inflation across countries might also be determined by structural external factors such as the exchange rate pass through. Understanding how the exchange rate fluctuations impact the price level across countries is important for the appropriate response of monetary policies. For this purpose, we will provide an analysis of several factors that may influence the exchange rate inflation, openness and the exchange rate regime. Taylor (2000) argues that there is a positive correlation between the rate of inflation and the reaction of prices in response to movements in the exchange rate. The rationale behind this is that firms are adjusting their prices more often in an inflationary environment due to more persistent costs. Therefore, the higher the inflation, the higher is the exchange rate pass through. The view put forth by Taylor (2000) finds support from the empirical literature. Choudri & Hakura (2006) test this hypothesis by employing a large dataset of both industrialized and developing countries. Their results confirm the positive correlation between the inflation rate and the pass through and, moreover, the relationship proves to be robust when they account for other macro variables. Similarly, the connection between the inflation rates and the exchange rate pass through proved to be significant in the study performed by Ca Zorzi et al (2007). Their evidence suggests a higher degree of pass through in emerging countries than in the developed ones. However, they find that the pass through is smaller in the emerging countries that experienced low inflation (such as Asia). The degree of openness is another determining factor of the exchange rate pass through. Openness is measured as the share of exports and imports to GDP or the share of imports to GDP. While it would be straightforward to assume that higher openness implies higher volatility in the exchange rate which later translates into higher prices, Romer (1993) finds a negative relationship between openness and inflation. Martinez & Iyer (2014) test this macroeconomic proposition for nine CEE countries. Their results are summarized in the Table 3.1. The negative correlation between openness and inflation has been confirmed for all countries, except Lithuania. The authors explain this contradiction by the potential difficulties faced by the monetary authorities to further decrease the inflation that was already low, given that openness has increased. According to Martinez & Iyer (2014), the coefficients

30 Theories and Factors of Inflation 31 for Bulgaria, Estonia, Latvia and Ukraine were not significant due to the restrictive fiscal measures implemented by the governments to bring down inflation. Table 3.1: Correlation between inflation and openness in nine Eastern European countries ( ). Source: Martinez & Iyer (2014). Jimborean (2011) estimates the exchange rate pass through to importer prices in NMS. The highest magnitude of the exchange rate pass through is found in Bulgaria (150 %), Estonia (107%) and Slovakia (95%), while the lowest pass through degree is captured in Lithuania (46%) and Czech Republic (48%). Despite that the paper does not focus on investigating the determinants of the exchange rate pass through, it presumes that countries with higher pass through coefficients are also the countries with high shares of imports to GDP. In contradiction with the aforementioned studies, Yamada & Bell (2012) find no statistical significance of the connection between trade openness and the inflation performance of Southeastern European countries. On the other hand, they conclude that the choice of the exchange rate regime influences inflation in the region. This factor is identified in the literature as one of the main determinants of the exchange rate pass through. Yamada & Bell (2012) argue that the choice of a fixed exchange rate regime by Bosnia and Herzegovina, Macedonia and Montenegro or of the IT regime (Albania) fosters a level of inflation significantly lower than under the flexible regime. The authors conclude that although Croatia has a flexible regime, it does not cause higher inflation in contrast to Serbia, which inflation is the highest in the region. Table 3.2: Exchange rate regimes in SEE countries. Source: Yamada & Bell (2012).

31 Theories and Factors of Inflation 32 In conclusion, the exchange rate pass through plays an essential role for the NMS. In the view of adopting the euro, these countries have to participate in the ERM II for at least two years. In these circumstances, the performance of NMS in effectively managing the exchange rate policies depends on the strength of the connection between the exchange rates and inflation. 3.2 Business cycle factors Inflation differentials across countries may also arise when their business cycles are not synchronized. To the extent that the output gap the difference between the actual and the potential output is a determinant of inflation in each country, the inflation differentials across CEEC shall therefore reflect different positions of countries in their business cycles. The importance of the output gap as an inflation determinant has been extensively discussed in the empirical literature. The study of Honohan & Lane (2003) documents a strong relationship between the output gap and inflation in the EMU. Similarly, Horvath & Koprnicka (2008) find a positive correlation between the output gap and inflation in NMS. Égert (2007) analyzes the impact of the cyclical factors on the inflation in the euro area and NMS by capturing this effect through the unit labor costs and the general government balance as a share of GDP. His results indicate that cyclical fluctuations are more robust in the euro area than in the CEE 10 countries. Stavrev (2009) decomposes inflation in NMS into common and country specific components and finds that the output gap, that enters the model as an idiosyncratic component, was a significant driver of inflation in NMS over Masso & Staehr (2005) find that the output gap exerts upward pressure on inflation in the Baltic States. Focusing on the demand side of the economy, the output gap can be explained by the growth of demand in excess of the GDP growth. One proxy for the demand pressure and output gap and is the real credit growth (ECB, 2003). Darvas & Szapáry (2008) argue that the rapid credit growth in NMS due to a sharp decline in the real interest rate can overheat the inflation. First, credit expansion causes wage growth that translates into higher prices and loss in competitiveness. They point out that the unit labor costs have increased at a fast pace in NMS, mostly in Latvia (Figure 3.1. Unit labor costs in manufacturing (annual growth in percent). Second, the rapid credit growth can increase the real house prices that in turn could further fuel the credit expansion and inflation the lower real interest rate due to higher inflation works in a pro cyclical way. The expansion of mortgage credit caused the real house prices in

32 Theories and Factors of Inflation 33 Estonia and Lithuania to increase by 30% and 20% during (Darvas & Szapáry, 2008). Figure 3.1. Unit labor costs in manufacturing (annual growth in percent). Source: Darvas & Szapáry (2008). Darvas & Szapáry (2008) suggest that the current account balance serves as a measure for the excess capacity in open economies. They argue that the current account deficits in some NMS reflect an increase in domestic demand rather than the loss in competitiveness due to wage growth that caused real effective exchange rate appreciation. The current account deficit as a proxy for cyclical determinants of inflation was also supported by Staehr (2009). The persistence of the cyclical inflation differentials depends mostly on the effectiveness of the equilibrating role of the competitiveness channel (ECB, 2012). The question, however, is whether the role of the real effective exchange rate through the loss of competitiveness can offset the expansionary effects of the real interest rate. For example, high inflation reduces the real interest rate and amplifies the inflation differentials, because of the pro cyclical interplay between inflation and real interest rate (Honohan & Lane, 2003). Meanwhile, high inflation causes real exchange rate appreciation that later exerts deflationary effects and brings down inflation differentials. However, the equilibrating effect of the real exchange rate occurs at a more gradual pace than the pro cyclical effect of the real interest rate and thus, the inflation differentials can persist for some time, especially if the wages are not flexible enough (Honohan & Lane, 2003).