DETERMINANTS OF THE DECISION TO IMPORT: A CROSS-COUNTRY COMPARISON Antonio Rodríguez Banco de España Joint work with Cristina Fernández, Coral García and Patry Tello Paper available at http://www.bde.es/f/webbde/ses/secciones/publicaciones/informesboletinesrevistas/boletineconomico/12/oct/files/art2e.pdf COMPNET. THE COMPETITIVENESS RESEARCH NETWORK Frankfurt 1th of December 212
OUTLINE OF THE PRESENTATION Motivation and literature review Dataset: Stylized facts Which are the determinants of importing at the firm-level? Conclusions 2
MOTIVATION (I) Increasing import dependence over the last few decades IMPORTS/GDP RATIO 2 1991 = 1 18 16 14 12 1 8 91 92 93 94 95 96 97 98 99 1 2 3 4 5 6 7 8 9 1 11 WORLD EURO AREA SOURCES: IMF and Eurostat. 3
MOTIVATION (II) and persistent differences across countries, NOT ONLY in the LEVEL of their import dependency 4
MOTIVATION (III) but ALSO in their imports STRUCTURE. STRUCTURE OF IMPORTS Nominal terms 1 8 % 6 4 2 2 211 2 211 2 211 2 211 France Germany Italy Spain SOURCE: Eurostat. SERVICES ENERGY NON-ENERGY INTERMEDIATE GOODS FINAL GOODS 5
A LITERATURE REVIEW Reasons for higher import dependency/differences in the import dependency and type of imported goods across countries International fragmentation of production/ Country specific factors/ Institutional factors/ Economic policy decisions/ Firms characteristics Importing goods and services: is it good or bad? It depends Augier et al. (29), Altomonte and Békés (29), Amiti and Konings (27), Keller (22) Only the more productive firms import: (Sunk) costs of the importing activity and the ability to exploit new and better inputs Altomonte and Békés (29), Vogel and Wagner (28), Muuls and Pisu (27) Set of firms characteristics that influence both on the propensity of importing and the type of goods purchased abroad Size, experience, human and technological capital, FDI/IO, EMN activity, financial constraints, spillover effects, importing hysteresis 6
WHAT THIS PAPER DOES Which firms characteristics increase the probability of being an importer? Are there differences across crountries? Which firms characteristics influence the type of imported goods? What lessons can we learn from the identified differences across countries? 7
THE DATASET: A DESCRIPTION Source of information: EFIGE survey Type of information: Firms characteristics (around 15 variables both qualitative and quantitative about the international activities, R&D, labour organization, financing and organizational activities and pricing behavior) Type of data: Manufacturing firm-level data. 28 cross-section data. Some variables also 27-29 average Dataset STRENGTHS Unique micro dataset: firm-level homogenous information for seven European countries It allows us to analyse the factors behind import propensity and the differences between countries Dataset WEAKNESSES Manufacturing firms with 1 employees or more; sample representativeness (energy); cross-section data (future waves); types of information (qualitative) This could bias the results: endogeneity problems/omitted variable bias/selfselection sample problems/correlations versus real causality 8
THE DATASET: MAIN STYLIZED FACTS (I) Profile of importing firms in the largest Euro Area economies RELATIVE WEIGHT OF IMPORTING FIRMS 6 5 4 3 2 1 % of total sample France Germany Italy Spain TOTAL IMPORTING FIRMS INTERMEDIATE GOODS SERVICES GEOGRAPHICAL BREAKDOWN OF IMPORTED INTERMEDIATE GOODS % of total intermediate goods importing firms 1 8 6 4 2 France Germany Italy Spain EU15 REST OF EU REST OF EUROPE INDIA AND CHINA REST OF ASIA USA AND CANADA CENTRAL ANDSOUTHAMERICA REST OF THE WORLD SOURCE: EU-EFIGE/Bruegel-UniCredit dataset. IMPORT INTENSITY (IMPORTED VALUE/NET TURNOVER) 2 16 12 8 4 % France Germany Italy Spain TOTAL IMPORTING FIRMS INTERMEDIATE GOODS SERVICES BREAKDOWN OF IMPORTED INTERMEDIATE GOODS BY COMPONENT % of total intermediate goods importing firms 1 8 6 4 2 RAW MATERIALS France Germany Italy Spain CUSTOMISED INTERMEDIATE GOODS STANDARD INTERMEDIATE GOODS Highest share of importers and import intensity in France, lowest in Germany Most diversified imports (both geo and by type of products) in France and Germany High dependency on raw materials in Spain and Italy 9
THE DATASET: MAIN STYLIZED FACTS (II) Main determinants of the decision to import IMPORTING vs NON-IMPORTING FIRMS. TOTAL SAMPLE 28 average 6 5 4 3 2 1 1 = Value of each variable in the case of non-importing firms Size Productivity Fixed capital per employee ratio Experience University graduat e ratio Product innovat ion SOURCES: EFIGE/Bruegel-UniCredit dataset and Amadeus. Process Group innovation membership Foreign group membership FDI International outsourcing Exporter Listed on stock exchange Sector-region spillovers a. The value of this variable indicates the percentage of firms having the characteristic in question (dummy variables). Internationalisation status, size, financial resources diversification 1
THE DATASET: MAIN STYLIZED FACTS (III) Main determinants of the decision to import: cross-country differences IMPORTING vs NON-IMPORTING FIRMS. SPAIN 28 average 1 25 9 8 7 6 5 4 3 2 1 1 = Value of each variable in the case of non-importing firms IMPORTING vs NON-IMPORTING FIRMS. FRANCE 28 average 1 25 9 8 7 6 5 4 3 2 1 1 = Value of each variable in the case of non-importing firms Size Productivity Group membership Foreign group membership FDI International outsourcing Listed on stock exchange IMPORTING vs NON-IMPORTING FIRMS. ITALY 28 average 1 25 9 8 7 6 5 4 3 2 1 1 = Value of each variable in the case of non-importing firms IMPORTING vs NON-IMPORTING FIRMS. GERMANY 28 average 1 25 9 8 7 6 5 4 3 2 1 1 = Value of each variable in the case of non-importing firms Size Productivity Group membership Foreign group membership FDI International outsourcing Listed on stock exchange SOURCES: EFIGE/ Bruegel-UniCredit dataset and Amadeus. a. The value of this variable indicates the percentage of firms having the characteristic in question (dummy variable). 11
DETERMINANTS OF THE DECISION TO IMPORT (I) The empirical strategy Econometric approach: a Probit is estimated to identify which variables could help to explain differences in the import propensity (country/sector specific factors and firms characteristics) Dependent variable: 1 if a firm imported in 28 and/or regularly in previous years, otherwise Control variables: Country and sector dummies Firms characteristics (size, experience, skills, R&D) Internationalisation status (foreign ownership, exporter, FDI/IO, import hysteresis) Industrial linkages (spillovers effects, group membership, IO, percentage of intermediate inputs on turnover) Financial constraints (stock exchange, bank debt) Presented results: coefficients reported are average marginal effects Dataset: EFIGE (Germany, France, Italy and Spain) 12
DETERMINANTS OF THE DECISION TO IMPORT (II) Dependent variable France (b) Italy (b) Spain (b) Size (c) University graduate ratio Process innovation Product innovation Foreign group membership FDI International outsourcing Exporter Sector-region spillovers SOURCE: EU-EFIGE/Bruegel-UniCredit dataset. IMPORTING FIRM [1] [2] [3] [4] [5].238***.239***.262***.243***.92*** (.15) (.14) (.14) (.14) (.17).39***.22***.65***.22* -.11 (.13) (.13) (.13) (.13) (.13).14***.14***.131***.114***.43*** (.14) (.14) (.14) (.13) (.14).17***.63***.6*** (.5) (.6) (.6).283***.18***.1** (.42) (.4) (.39).5***.45***.4*** (.1) (.1) (.9).12***.74***.74*** (.1) (.1) (.1).11***.79*** (.23) (.23).146***.136*** (.25) (.24).199***.195*** (.26) (.25).236***.216*** (.9) (.1).582*** (.33) a. Average marginal effects are reported. Standard deviations are in brackets. *, **, *** denote statistical significance at 1%, 5% and 1%, respectively. Full table available in the paper. b. Reference taken is Germany. c. As natural logarithm. 13
DETERMINANTS OF THE DECISION TO IMPORT (III) The empirical strategy Econometric approach: a Probit (Heckman) is estimated to identify which variables could help to explain differences in the type of goods imported Dependent variable: 1 if a firm imported ONLY one type of good, otherwise Control variables: Country and sector dummies Firms characteristics (experience, skills, R&D) Internationalisation status (foreign ownership, exporter, FDI/IO) Industrial linkages (spillovers effects, group membership) Financial constraints (stock exchange) Presented results : coefficients reported are average marginal effects Dataset: EFIGE (Germany, France, Italy and Spain) 14
DETERMINANTS OF THE DECISION TO IMPORT (IV) Dependent variable IMPORTING FIRM ONLY RAW MATERIALS France (b) Italy (b) Spain (b) Size University graduate ratio Process innovation Product innovation Foreign group membership FDI International outsourcing Exporter Sector-region spillovers SOURCE: EU-EFIGE/Bruegel-UniCredit dataset. Most efficient firms show a LOWER propensity of importing ONLY raw materials (in general, the opposite to other intermediate goods) Country dummies are significant ONLY STANDARD INTERMEDIATE GOODS ONLY CUSTOMISED INTERMEDIATE GOODS [1] [2] [3] [4].92*** -.54** -.17*** -.46*** (.17) (.27) (.25) (.16) -.11.31*** -.51*** -.15* (.13) (.4) (.14) (.8).43***.187*** -.14 -.38*** (.14) (.38) (.1) (.14).6*** (.6).1** -.1.61**.37* (.39) (.65) (.3) (.22).4*** -.41*** -.4.2 (.9) (.14) (.7) (.5).74*** -.85***.2.9 (.1) (.15) (.8) (.5).79*** -.26.13.17* (.23) (.28) (.14) (.1).136*** -.97*** -.19. (.24) (.29) (.15) (.11).195*** -.194*** -.11.28*** (.25) (.33) (.17) (.1).216*** -.96*** -.1.11 (.1) (.26) (.13) (.8).582*** -.156**.65***.25 (.33) (.74) (.21) (.18) a. Average marginal effects are reported. *, **, *** denote statistical significance at 1%, 5% and 1%, respectively. Full table in the paper. b. Reference taken is Germany. c. As natural logarithm. 15
CONCLUSIONS There are important differences in the import dependency across the four largest Euro Area countries and also in the type of products imported Firms characteristics only explain part of these differences Country-specific characteristics are relevant as well as spillover effects Aggregate figures of import dependency don t have to be interpreted as a negative indicator of a country s competitiveness. It all depends on the type of goods imported. Fostering human and technological capital would help to take full advantage of international trade on imported inputs 16
THANK YOU FOR YOUR ATTENTION DIRECTORATE GENERAL OF ECONOMICS, STATISTICS AND RESEARCH
THE DATASET: MAIN STYLIZED FACTS (IV) Main determinants of the decision to import: cross-country differences. Importing firms compared to sample average IMPORTING FIRMS. SPAIN vs SAMPLE 28 average 1 = Value of each variable in the case of the sample importing firms 14 IMPORTING FIRMS. ITALY vs SAMPLE 28 average 1 = Value of each variable in the case of the sample importing firms 14 12 1 8 6 4 2 IMPORTING FIRMS. FRANCE vs SAMPLE 28 average 1 = Value of each variable in the case of the sample importing firms 18 16 14 12 1 8 6 4 2 Size Productivity Group FDI membership Foreign group membership International outsourcing SOURCES: EFIGE/ Bruegel-UniCredit dataset and Amadeus. Listed on stock exchange 12 1 8 6 4 2 IMPORTING FIRMS. GERMANY vs SAMPLE 28 average 1 = Value of each variable in the case of the sample importing firms 18 16 14 12 1 8 6 4 2 Size Productivity Group Foreign FDI International membership group outsourcing membership Listed on stock exchange a. The value of this variable indicates the percentage of firms having the characteristic in question (dummy variables). 18