Business Cycle Synchronisation in Europe: Core vs. Periphery

Transcription

1 Business Cycle Synchronisation in Europe: Core vs. Periphery Ansgar Belke, Clemens Domnick, Daniel Gros

2 Differences in Economic Activity: I unemployment rate q1 2000q1 2005q1 2010q1 2015q1 Core Periphery Notes: Arithmetic average with Core = {Austria, Germany, Netherlands, France, Finland} and Periphery={Portugal, Italy, Ireland, Spain, Greece}. Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

3 Differences in Economic Activity: II GDP (1999Q1=100) q1 2005q1 2010q1 2015q1 Core Notes: Core = Austria, Germany, Netherlands, France, Finland Periphery= Portugal, Italy, Ireland, Spain, Greece Periphery Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

4 Outline 1 Motivation and Introduction 2 Empirical Framework 3 Estimation Results 4 Further Research and Conclusion Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

5 Why is BC synchronization important? Most important critera in Optimum Currency Area theory Relevant for monetary policy Stabilization policy Asymmetric shock propagation - (Altavilla, 2004) Smoother monetary policy transmission Relevant for governance of EMU Do we need further stabilization mechanisms? (De Grauwe and Ji, 2016) "Cost" of EMU depends on business cycle (BC) synchronization Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

6 This Paper What is variation in output co-movement between core and peripheral countries? (Re-)assessment of BC synchronization in the Eurozone Sample Cross-Section: Focus on Core-Periphery split Time Period: Focus on BC in period after the great financial crisis High-frequency Index by Cerqueira and Martin (2009) Different Methods: Parametric and nonparametric regressions as flexible framework Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

7 Core-Periphery in the Literature Academia, among others: Current Account Balance (De Santis and Cesaroni, 2016) Fiscal Policy (Blanchard et al., 2016) Competitiveness (Wortmann and Stahl, 2016) How to define "core" and "periphery"? Geographic proximity? Economic Policy? Cluster analysis? Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

8 Literature Review Before the EMU: Dickerson et al. (1998), Angeloni and Dedola (1999) Introduction of EMU: verdict unclear Theoretically: endogeneous OCA (Frankel and Rose, 1997) vs. Specialization (Krugman, 1991) Increase in synchronization: Furceri and Karras (2008), Gayer (2007) No change or decrease: Weyerstrass et al. (2011), Camacho et al. (2006) Differences core vs. periphery: Lehwald (2013), Gouveia and Correia (2008) Crisis-Period: Literature scarce, but points to decoupling of periphery Synchronization lower in EMU (Gächter et al., 2012, Grigoas and Stanciu, 2016) GFC (2007) triggered lower co-movement for Greece, Ireland, Portugal and Spain (Degiannakis et al., 2014) Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

9 Different Approaches Compare non-overlapping correlation coefficients (Furceri and Karras, 2008) Concordance Indices (Grigoas and Stanciu, 2016) Dynamic Factor Models (Lehwald, 2013) Dynamic Correlation (Weyerstrass et. al, 2011) Wavelet Analysis (Aguiar-Conraria and Soares, 2011) Local Polynomial Regressions (Cerqueira, 2013) Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

10 Data Quarterly Real GDP, seasonally adjusted (OECD) EA-12 countries plus Denmark, Norway, Switzerland and Sweden, 1970Q1-2015Q4 Cyclical component with Hodrick-Prescott filter, "growth cycle" (Lucas, 1977) Correlation Index ρ ijt pioneered by Cerqueira and Martin (2009) and Cerqueira (2013) Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

11 Data: Correlation Index ρ ijt Intuition: Transformation of "classical" correlation coefficient into a high(er) frequency measure Details Symmetric around 0 (unbounded) Well-suited for statistical analysis (no autocorrelation) Gains traction in BC literature (Bierbäumer-Polly et al. (2016), Gächter and Riedl (2014)) Example: Detrend French and Italian GDP (HP-Filter) Calculate correlation index for each quarter Result: one dyadic (France-Italy), quarter-specific observation belonging to "Core-Periphery" Cluster Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

12 Empirical Framework 1 Classical Correlation Coefficients 2 Panel Regressions For (1) and (2) ad-hoc definition of crisis period (starting 2008Q1), following Bekiros et al. (2015) 3 Local Polynomial Regressions (LPR) Non-parametric approach Flexible dynamic analysis to detect patterns in the data (1) and (2) ad-hoc approaches, (3) dynamic analysis For (2) and (3) correlation index as dependent variable Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

13 (1) Simple Correlation Coefficients Pre-Crisis Periphery Core non-emu EA-12 Periphery 1 Core non-emu EA Crisis Periphery 1 Core non-emu EA Notes: Pre-crisis ranging from 1999Q1-2007Q4, "crisis" period from 2008Q1-2015Q4. ***/**/* indicate significance at the 1%/5%/10% level. Core = Germany, France, Netherlands, Finland, Austria Periphery = Italy, Ireland, Greece, Portugal, Spain non-emu = Sweden, Switzerland, Norway, Denmark Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

14 (2) Panel Estimation: Country-Pair View ρ ijt = β 0 + β 1 Crisis t + β 2 Cluster ij Crisis t + δ i + γ j + α ij + ω t + ɛ ijt ρ ijt = Correlation Index β 1 Crisis t = dummy variable 1 in period 2008Q1-2015Q4 β 2 Cluster ij Crisis t = interaction term between Cluster and Crisis Cluster ij = dummy variable equaling 1 for country-pairs belonging to specific Cluster (Core, Periphery, non-emu) Country (δ i, γ j ), Country-Pair (α ij ) and Time FE (ω t ) Quarterly data (1999Q1-2015Q4) Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

15 (2) Panel Estimation: Results (1) (2) (3) Crisis (0.140) (0.123) (0.122) Crisis Core-Per. Crisis Per. - Per. Crisis Core - Core (0.0604) (0.0880) (0.0880) Crisis Cor. - EA (0.0518) Crisis Per. - EA (0.0515) Crisis Core-NonEMU (0.0680) Crisis Per-NonEMU (0.0638) N R Time FE Country FE Bilateral FE Sample Mean ρ Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

16 (3) Local Polynomial Regressions (LPR) Motivation: ρ ijt = f (t) + u ijt What is the shape of the regression function? No assumptions about functional form Separate signal from noise Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

17 (3) LPR: Core and Periphery Core Periphery Periphery Periphery kernel = epanechnikov, degree = 1, bandwidth = 6.39, pwidth = 9.59 kernel = epanechnikov, degree = 1, bandwidth = 5.59, pwidth = 8.38 Core Core kernel = epanechnikov, degree = 1, bandwidth = 9.47, pwidth = Robustness Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

18 (3) LPR: Core/Periphery vs. Non-EMU non EMU Core non EMU Periphery kernel = epanechnikov, degree = 1, bandwidth = 7.5, pwidth = kernel = epanechnikov, degree = 1, bandwidth = 5.98, pwidth = 8.97 non EMU EA kernel = epanechnikov, degree = 1, bandwidth = 5.81, pwidth = 8.71 Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

19 Conclusion 1 Core and periphery are less aligned after GFC 2 Core in itself stronger, periphery shows less correlation with itself 3 Output co-movement fell more vis-à-vis non-emu countries for periphery than for core Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

20 Further Research and Limitations What are the channels? 1 Trade? 2 Financial Flows? 3 Nominal Wage Costs? What does this mean for economic governance of the EMU? Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

21 THANK YOU! Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

22 Literature Angeloni, I. and Dedola, L. (1999). From the ERM to the euro: new evidence on economic and policy convergence among EU countries. ECB Working Paper No. 4. Aguiar-Conraria, L. and Soares, M. J. (2011). Business cycle synchronization and the Euro: A wavelet analysis. Journal of Macroeconomics, 33 (3), Altavilla, C. (2004). Do EMU members share the same business cycle? Journal of Common Market Studies, 42 (5), Bekiros, S., Nguyen, D. K., Uddin, G. S. and Sjö, B. (2015). Business cycle (de) synchronization in the aftermath of the global financial crisis: implications for the Euro area. Studies in Nonlinear Dynamics and Econometrics, 19 (5), Camacho, M., Perez-Quiros, G. and Saiz, L. (2006). Are European business cycles close enough to be just one? Journal of Economic Dynamics and Control, 30 (9 10), Cerqueira, P. A. (2013). A closer look at the world business cycle synchronization. IEEP, 10 (3), Cerqueira, P. A. and Martins, R. (2009). Measuring the determinants of business cycle synchronization using a panel approach. Economics Letters, 102 (2), De Grauwe, P. and Ji, Y. (2016). Flexibility versus Stability. A difficult trade-off in the Eurozone. CEPR Discussion Paper De Santis, R. and Cesaroni, T. (2016). Current Account Core-Periphery Dualism in the EMU. The World Economy. Degiannakis, S., Duffy, D. and Filis, G. (2014). Business Cycle Synchronization in EU: A Time-Varying Approach. Scottish Journal of Political Economy, 61 (4), Dickerson, A. P., Gibson, H. D. and Tsakalotos, E. (1998). Business cycle correspondence in the EU. Empirica, 25 (1), Frankel, J. A. and Rose, A. K. (1997). Is EMU more justifiable ex post than ex ante? EER, 41 (3), Furceri, D. and Karras, G. (2008). Business-cycle synchronization in the EMU. Applied Economics, 40 (12), Gächter, M. and Riedl, A. (2014). One money, one cycle? The EMU experience. Journal of Macroeconomics, 42, Gächter, M., Riedl, A. and Ritzberger-Grünwald, D. (2012). Business Cycle Synchronization in the Euro Area and the Impact of the Financial Crisis. Monetary Policy and the Economy, 2, Gouveia, S. and Correia, L. (2008). Business cycle synchronisation in the Euro area: the case of small countries. IEEP, 5 (1-2), Grigoras, V. and Eusignia Stanciu, I. (2016). New Evidence on the (De)Synchronisation of Business Cycles: Reshaping the European Business Cycle. International Economics. Krugman, P. R. (1991). Geography and trade. MIT press. Lehwald, S. (2013). Has the Euro changed business cycle synchronization? Evidence from the core and the periphery. Empirica, 40 (4), Weyerstrass, K., van Aarle, B., Kappler, M. and Seymen, A. (2011). Business cycle synchronisation with (in) the EURO area: in search of a euro effect. Open Economies Review, 22 (3), Wortmann, M. and Stahl, M. (2016). One Size Fits Some: A Reassessment of EMU s Core periphery Framework. Journal of Economic Integration, 31 (2), Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

23 Appendix Construction of Correlation Index "Bounded" Correlation Index (between 1 and 3-2T) ρ b ijt = T g it g i Tt=1 (g it g i ) 2 1 T 2 g jt g j Tt=1 (g jt g j ) 2 "Fisher Transformation" for unbounded Index (centered around 0) ρ ijt = 1 2 ln 1 + ρb ijt 1 ρ b ijt 2T 3 Return to Presentation Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

24 Appendix LPR: Different Definitions for Periphery Core Per without Greece Core Periphery Core Per without Italy Return to Presentation Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

25 Appendix Core vs. Periphery, Aggregate View Correlation Index (Fisher Transformation) kernel = epanechnikov, degree = 1, bandwidth = 7.48, pwidth = Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

26 Appendix Core vs. Periphery, Aggregate View, Smoothed vs. Original Value LPR Correlation Index Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

27 Appendix Contemporaneous Effect of EA-12 Cycle on national Cycle: Periphery Greece Italy Spain Ireland Portugal EA (0.268) (0.0389) (0.0915) (0.160) (0.0809) R Notes: Dependent variable is the quarterly (HP-filtered) cyclical component of each peripheral country. Sample from 1999Q1-2015Q4. Robust standard errors in parenthesis. ***/**/* indicate significance at the 1%/5%/10% level. Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

28 Appendix Contemporaneous Effect of EA-12 Cycle on national Cycle: Core Finland France Germany Austria Netherlands EA (0.0785) (0.0332) (0.0505) (0.0494) (0.0607) R Notes: Dependent variable is the quarterly (HP-filtered) cyclical component of each core country. Sample from 1999Q1-2015Q4. Robust standard errors in parenthesis. ***/**/* indicate significance at the 1%/5%/10% level. Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery

29 Appendix Contemporaneous Effect of EA-12 Cycle on national Cycle: non-emu Sweden Denmark Norway Switzerland EA (0.0936) (0.0648) (0.0840) (0.0515) R Notes: Dependent variable is the quarterly (HP-filtered) cyclical component of each core country. Sample from 1999Q1-2015Q4. Robust standard errors in parenthesis. ***/**/* indicate significance at the 1%/5%/10% level. Belke, Gros, Domnick (U. Duisburg-Essen) BC: Core vs. Periphery