Beggar-Thy-Neighbor or Beneficial Spillover: Effect of Exchange Rates on GVC Trade

Transcription

1 Beggar-Thy-Neighbor or Beneficial Spillover: Effect of Exchange Rates on GVC Trade 5 th IMF WB WTO Joint Trade Research Workshop November 30, 2016 Gee Hee Hong (IMF) (with Kevin Cheng, Dulani Seneviratne, Rachel van Elkan (all IMF)) 1

2 Motivation Conventional Wisdom on Exchange Rate Policy and Trade: Beggar-Thy-Neighbor : Competitive devaluation or a weaker currency spurs exports and economic growth of depreciating country at the expense of other countries However, not so clear with GVC production structure as: A country s imports and exports are not substitutes, but complements in GVC structure 2

3 International Fragmentation of Production Process 3

4 Main Questions GVC Feature: Production and exports require imports of intermediate goods. Key Question: Q1. How do GVC-related Exports and Imports respond to changes in exchange rates? Q2. Does Beggar-Thy-Neighbor Effect hold in the GVC trade structure? Q3. What determines the (value-added) exchange rate elasticities of GVC-related trade? 4

5 Value-added of this paper Beggar-Thy-Neighbor or Beneficial Spillovers? Provide a theoretical underpinning for result that a real appreciation can: (i) reduce exports of DVA; and (i) reduce imports of FVA. Spill-forward and spill-back. Position in the production chain is not important; rather, what matters is how large is your contribution! Empirically, using DVA and FVA components separately, we find that Real appreciation reduces exports of DVA (a conventional result ) AND reduces imports of FVA (contrary to traditional theory). A larger foreign contribution to gross GVC-related exports dampens the response of DVA and FVA to REER changes. 5

6 Model Set-Up 3 Country Model of a Snake GVC Structure: Gross output: Qi (i=1,2,3), where Q3 is final product Domestic Value Added: Vi (i=1,2,3) Foreign Value Added (Backward Linkage): Q(i 1) CES Production Function (w/ common elasticity of substitution) Country 1 Country 2 Country 3 V1=Q1 Country 1 Export: Q1 FVA: Q1 DVA: V2 Country 2 Export: Q2 FVA: Q2 DVA: V3 Final Output:Q3 6

7 Model Production Technology (CES Technology) is the share of country i s contribution in gross output of j, : Price of gross output of country i : Price of value-added of country i in own currency : : in country j s currency Final Good: 7

8 Model: Real Depreciation and DVA Proposition (Standard Beggar-Thy-Neighbor). A real depreciation will increase domestic value added of GVC related gross output and exports. The elasticity is smaller if import content in gross output is larger. Proof. Profit Maximization: Log Linearization: Price Index: Rewriting (9), and. 8

9 Model: Real Depreciation and FVA Proposition (Beneficial Spillover). If a country s own value added contribution to gross output is sufficiently large, a real depreciation in that country will raise the amount of value added by other countries in the chain. Therefore, a downstream country that experiences a real depreciation will import more from the upstream country, even though the upstream country s real exchange rate has appreciated beneficial spillback. Proof. Profit Maximization: Scale Effect: Own-price effect Cross-price effect Scale effect 9

10 Model: Real Depreciation and FVA (Continued..) Proof (continued.) : ). 10

11 Model: Real Depreciation and FVA (Continued..) Corollary (beneficial spill forward). If the contribution of the upstream country s domestic value-added to the final product is sufficiently large, a real depreciation in the upstream country may raise the value-added produced by the downstream country, even though the downstream country has experienced a relative real appreciation. 11

12 Data Description Data Set: OECD-WTO TiVa Database from (Duval et al. (2015)) Key Variable Description: Construction of REER : (Bayoumi et al. (2005), weights based on DVA in exports) Share of FVA in GVC-Related Exports: 12

13 Empirical Strategy Baseline Regression: Novelty: Distinguishes GVC related trade and non GVC related trade Use VA weighted REER 13

14 Results Table. Response of Trade to the Real Exchange Rate VARIABLES (1) (2) (3) (4) (5) (6) Gross Trade Other Trade GVC-related exports Gross Gross Imports Exports Imports Exports Imports Exports (FVA DD ) (DVA Non-GVC ) (FVA X ) (DVA GVC ) Lagged Log (REER-Value-added-based) * 1.235** ** *** *** (0.663) (-1.841) (2.287) (-2.378) (-2.822) (-3.527) lagged log(reer) x lagged (FVA/DVA+FVA) ** 0.027*** 0.026*** (-0.443) (2.150) (3.166) (3.330) Lagged Log (Demand) 0.709*** 1.110*** 0.884*** 0.795*** 1.108*** 0.758*** (4.716) (9.257) (3.089) (6.446) (5.961) (4.470) Time FE Y Y Y Y Y Y Country FE Y Y Y Y Y Y Additional controls Y Y Y Y Y Y Clustering Country level Country level Country level Country level Country level Country level Observations R-squared Source: IMF staff estimates. N ote: REER stands for real effectiv e ex change rate. Specification log (Ex ports [Imports] v olume)c,t = αt +αc + α1log(reer)c,t-1 +α2interaction term + α3log(demand)c[w],t-1 + αilog(controls)c,t + ɛt. Additional controls included in the specifications are log of real stock of FDI, FVA share, tariffs, output gap. Demand is prox ied by GDP. Robust t-statistics in parentheses. *** p<0.01, ** p<0.05, * p<0.1 14

15 Results Table. Exchange Rate Competitiveness in GVCs (1) (2) VARIABLES Gross imports Gross exports Lagged Log (REER-Value-added-based) * (0.663) (-1.841) lagged log(reer) x lagged (FVA/DVA+FVA) ** (-0.443) (2.150) Lagged Log (Demand) 0.709*** 1.110*** (4.716) (9.257) 15

16 Results Table. Exchange Rate Competitiveness in GVCs (3) (4) VARIABLES Non-GVC imports Non-GVC exports Lagged Log (REER-Value-added-based) 1.235*** *** (2.287) (-2.378) lagged log(reer) x lagged (FVA/DVA+FVA) n/a n/a Lagged Log (Demand) 0.884*** 0.795*** (3.089) (6.446) 16

17 Results Table. Exchange Rate Competitiveness in GVCs (5) (6) VARIABLES GVC imports (FVA) GVC exports (DVA) Lagged Log (REER-Value-added-based) *** *** (-2.822) (-3.527) lagged log(reer) x lagged (FVA/DVA+FVA) 0.027*** 0.026*** (3.166) (3.330) Lagged Log (Demand) 1.108*** 0.758*** (5.961) (4.470) 17

18 Impact on GVC-trade elasticities GVC Trade Elasticities (with 90 percent Confidence Intervals) GVC Exports (GVC-related DVA) 1.0 GVC Imports (FVA) GVC Trade Elasticity GVC Trade Elasticity FVA Share in GVCs (percent) FVA Share in GVCs (percent) 18

19 Impact on (overall) Elasticities Total DVA Exports Elasticity given a Change in DVAgvc Weight 1.0 s=25 s=50 s=75 TotalDVA export elasticity Weight of GVC-related DVA in total DVA (range=0-1)

20 Impact on Net Trade Balance Frequency Distribution: Impact on the Net Trade Balance of a 10 Percent Real Depreciation (Based on actual FVA shares and GVC trade weights in the sample of countries) Frequency Trade Balance (in percent of GDP)

21 Conclusion GVCs alter the responsiveness of trade to exchange rate changes; Responsiveness depends on a country s contribution to the final GVC product. Higher FVA in your exports dampening Complicate assessments of impact of exchange rate changes on trade. Beggar-Thy-Neighbor or Beneficial Spillover? -> A real depreciation may not be a bad thing for your GVC trading partner. 21

22 Supplementary Slides 22

23 Breakdown of Gross-trade data Gross Exports Gross Imports DVA DVA related to GVCs FVA Traditional Imports (not re-exported) DVA related to traditional exports 23

24 Decomposition Decomposition of Gross Exports and Imports, 1995 vs (In percent of GDP) 70 Non-GVC-related imports Non-GVC-related DVA in exports GVC-related DVA in exports GVC-related FVA in exports DVA FVA Asia ASEAN Euro area Full sample Asia ASEAN Euro area Full sample Gross Exports Imports Sources: OECD-WTO TiVA database; and IMF staff estimates. 24

25 Data Description Decomposition of Gross Exports and Imports, 1995 vs (In percent of GDP) 70 Non-GVC-related imports Non-GVC-related DVA in exports GVC-related DVA in exports GVC-related FVA in exports DVA FVA Asia ASEAN Euro area Full sample Asia ASEAN Euro area Full sample Gross Exports Imports Sources: OECD-WTO TiVA database; and IMF staff estimates. 25