International Inflation Spillovers Through Input Linkages e 1 Raphael A. Auer Andrei A. Levchenko Philip Saur´ BIS, U Michigan, SNB September 30, 2016 1 The views expressed here are those of the authors and do not necessarily reflect those of the BIS or the SNB.
Motivation Inflation is highly synchronized across countries Important to know why Inflation forecasting Monetary policy and its international coordination Currency unions Hypothesis: inflation comoves across countries due to input linkages. Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 2/27
This Paper Assess empirically the role of cross-border input linkages in inflation synchronization Main idea: PPI c = γ c , e × � � PPI e + � C c A unique database that combines sectoral PPI inflation for 30 countries and 17 sectors with the World Input-Output matrix (WIOD) Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 3/27
Preview of Results 1. Transmission of hypothetical shocks A 1% global inflation shock raises average country PPI by 0.23% Shocks to large individual countries matter for their closest smaller trading partners E.g. Canada wrt USA, Hungary wrt Germany have an elasticity of around 0.1 2. Tracking the Global Component of Inflation : input linkages account for about half of observed inflation comovement. A single common factor accounts for about 50% of the variance of PPI inflation, but less than 50% of the variance of the underlying cost shocks Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 4/27
Preview of Results (continued) 3. Mechanisms : Exchange rate movements do not contribute to global price comovment: β = 0 . 3 → linkages explain only about 15-20% of comovement Heterogeneity in input linkages matters somewhat, but most of the effect is driven by their average level Both the comovement of PPI, and the contribution of linkages to comovement, are driven by common sectoral shocks 4. Distribution : inflation behavior exhibits fat tails (relative to a normal distribution). Input linkages preserve fat-tailedness of underlying shocks, rather than average them out. Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 5/27
Literature International inflation synchronization: Monacelli and Sala (2009), Burstein and Jaimovich (2012), Andrade and Zachariadis (2015), and Beck, Hubrich and Marcellino (2015); Ciccarelli and Mojon (2010), Mumtaz and Surico (2009, 2012) and Mumtaz, Simonelli and Surico (2011); Borio and Filardo (2007) and Bianchi and Civelli (2015); Input linkages and international relative prices: Bems and Johnson (2012, 2015) and Patel, Wang and Wei (2014) International business cycle comovement through input linkages: Kose and Yi (2006), Burstein, Kurz and Tesar (2008), di Giovanni and Levchenko (2010), Johnson (2014) Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 6/27
Cost Function and PPI C countries, indexed by c and e , and S sectors, indexed by u and s . Sector u in country c cost function W c , u , t = W ( C c , u , t , p c , u , t ) , with p c , u , t = ( p c , u , e , s , t ) e , s In changes: � � c , u , t − 1 � W c , u , t ≈ γ C γ I C c , u , t + c , u , e , s , t − 1 � p c , u , e , s , t e , s Assumptions: 1. � PPI c , u , t = � W c , u , t (constant markups) � � W e , s , t + � � p c , u , e , s , t = β I 2. Imported input prices: � E c , e , t c , u , e , s Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 7/27
Recovering the Cost Shocks Cost shock recovered directly � � � 1 � � PPI e , s , t + � � β I c , u , e , s γ I C c , u , t = PPI c , u , t − E c , e , t c , u , e , s , t − 1 γ C c , u , t − 1 e ∈ C , s ∈ S 12-month changes ( X = { PPI , C } ): 11 � � (1 + � X 12 c , u , t = X c , u , t − τ ) − 1 τ =0 Aggregated: � � ω c , u � X 12 c , t = X 12 c , u , t u ∈ S Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 8/27
Data PPI data: National statistical offices (Eurostat, BLS, StatCan, ...) Country-specific product classification Frequency: monthly Cross-border trade and output data: World Input-Output database (WIOD) Final sample: 17 sectors, 30 countries + ROW; 1995m1-2011m12 Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 9/27
Countries and Sectors Country Code Sector Australia AUS Agriculture, Hunting, Forestry, and Fishing Austria AUT Basic Metals and Fabricated Metal Belgium BEL Chemicals and Chemical Products Bulgaria BGR Coke, Refined Petroleum and Nuclear F.. Canada CAN Electrical and Optical Equipment China CHN Electricity, Gas and Water Supply Czech Republic CZE Food, Beverages and Tobacco Denmark DNK Leather, Leather and Footwear Finland FIN Machinery, Nec France FRA Manufacturing, Nec; Recycling Germany DEU Mining and Quarrying Greece GRC Other Non-Metallic Mineral Hungary HUN Pulp, Paper, Paper , Printing and Pub.. Ireland IRL Rubber and Plastics Italy ITA Textiles and Textile Products Japan JPN Transport Equipment Korea KOR Wood and Products of Wood and Cork Lithuania LTU Mexico MEX Netherlands NLD Poland POL Portugal PRT Rest of the World ROW Romania ROM Russian Federation RUS Slovenia SVN Spain ESP Sweden SWE Taiwan, POC TWN United Kingdom GBR United States USA Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 10/27
Imported Input Use by Country Auer, Levchenko, and Saur´ Share 0 .2 .4 .6 .8 e BEL NLD HUN SVN LTU IRL DNK International Inflation Spillovers AUT TWN CZE SWE DEU BGR PRT CAN GRC MEX POL GBR FRA FIN ESP ITA September 30, 2016 ROM KOR AUS USA ROW JPN CHN RUS 11/27
Result 1. Transmission of Hypothetical Shocks Through IO Matrix Equilibrium � PPI : PPI = ( I − Γ ′ ) − 1 � � C Hypothetical shock to an individual country: � � ′ � C s , 1 · · · � � C = 0 · · · 0 C s , J 0 · · · 0 Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 12/27
1% US Inflation USA MEX CAN TWN KOR IRL CHN BEL NLD HUN DEU JPN GBR FRA CZE SVN AUT AUS SWE FIN ESP BGR DNK PRT POL ITA ROM GRC LTU RUS 0 .2 .4 .6 .8 1 %points ∆ PPI Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 13/27
Each Country on Each Country 0.14 0.12 impact (destination shock/source shock) 0.1 0.08 0.06 0.04 0.02 DEU CHN RUS USA FRA NLD BEL ITA HUN GBR CZE JPN SVN BEL TWN ESP LTU AUT POL BGR AUT SWE CAN NLD KOR KOR SWE POL AUS DEU PRT CZE FIN TWN DNK DNK IRL FIN MEX FRA HUN ROM MEX CAN ROM ESP IRL GRC GBR PRT ITA source BGR CHN SVN USA GRC JPN destination LTU AUS RUS Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 14/27
1% Worldwide Inflation BEL HUN CZE SVN TWN LTU AUT BGR SWE NLD KOR POL DEU PRT FIN DNK IRL MEX FRA ROM CAN ESP GRC GBR ITA CHN USA JPN AUS RUS 0 .1 .2 .3 .4 %points ∆ PPI Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 15/27
10% Change in Energy Prices LTU RUS BGR GRC TWN HUN BEL NLD KOR JPN FIN USA FRA CAN POL ROM MEX ESP DEU AUS ITA SWE CZE CHN PRT AUT DNK GBR IRL SVN 0 1 2 3 4 %points ∆ PPI Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 16/27
Decomposition Direct and Indirect Effects 10% PPI Inflation Shock in China 8 Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 17/27
Result 2. Tracking the Global Component of Inflation Metrics of synchronization 1. R 2 of the country’s � PPI ( � C ) on world average � PPI ( � C ) (Ciccarelli-Mojon, 2010) Share of the variance explained by a world factor: X c , t = λ c F t + ǫ c , t Var ( λ c F t ) Var ( X c , t ) 2. Static factor 3. Dynamic factor Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 18/27
Result 2: Synchronization, Country-Level, β = 1 Input linkages account for at least half of observed inflation comovement. Panel A: R 2 Panel B: Static Factor Panel C: Dynamic Factor � � � � � � PPI 12 c , t C 12 c , t PPI 12 c , t C 12 c , t PPI 12 c , t C 12 c , t Mean 0.385 0.172 0.455 0.268 0.444 0.235 Median 0.365 0.110 0.506 0.286 0.500 0.181 Min 0.006 0.000 0.000 0.002 0.002 0.002 Max 0.776 0.527 0.931 0.812 0.916 0.761 Recall: PPI c = β × γ c , s × � � PPI s + � C c Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 19/27
Variance Shares: Cross-Country Heterogeneities 1 .8 Var( λ c F t )/Var(X ct ) .6 .4 .2 0 ESP DEU ITA NLD BEL JPN FRA LTU FIN CHN CAN AUS AUT USA GBR PRT SWE TWN BGR CZE DNK RUS POL GRC IRL HUN MEX KOR SVN ROM PPI C Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 20/27
Result 3: Mechanisms 1. Exchange rate movements 2. Heterogeneity in linkages 3. Sectoral vs. global shocks Auer, Levchenko, and Saur´ e International Inflation Spillovers September 30, 2016 21/27
Recommend
More recommend
