## _wp12236

## Source details

**Canonical URL:** [_wp12236](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12236.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12236.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12236.pdf.json)

---

### Overview
- After its first decade, the euro area faced severe pressure triggered by fiscal adjustments in Greece that spread to Ireland and Portugal and later threatened the eurozone’s existence as Italy and Spain experienced funding pressures.
- Crisis drivers emphasized in the chapter:
  - Intertwined public debt and banking sector fragilities, weak growth prospects, and substantial gross and net external liabilities.
  - Net external liabilities close to 100 percent of GDP in Greece, Ireland, Portugal, and Spain (end-2010).
- Focus: the external dimension of the euro area crisis, emphasizing five largest “net debtors”: Greece, Ireland, Italy, Portugal, and Spain.

### Main findings on causes of imbalances
- Large current account imbalances of individual euro area countries reflected:
  - The asymmetric impact of trade shocks originating outside the euro area.
  - Sustained cheap financing from core euro area countries to the largest net debtors.
- External drivers documented:
  - Rise of China increased demand for machinery and equipment exported by Germany while displacing exports from euro-area debtor countries.
  - Terms of trade shock from higher oil prices contributed to rising trade deficits and boosted demand for German machinery and equipment.
  - German outward integration (production platforms in emerging Europe) boosted emerging Europe competitiveness and exports to debtor countries.
- Financing and exchange rate dynamics:
  - Continued easy financing allowed deficit countries to sustain appreciating real effective exchange rates; nominal appreciation of the euro also contributed, delaying adjustment.

### Distinct contributions relative to existing literature
- Emphasis on trade linkages and relative price dynamics between euro area countries and rest of world:
  - Between 2000 and 2009 the lion share of REER appreciation was due to the nominal appreciation of the euro vis-à-vis other currencies.
  - Econometric evidence: euro appreciation adversely affected export performance of debtor countries over and above the average euro-area effect.
  - Bilateral trade contributions documented: China, oil exporters, and emerging Europe to Germany’s surplus and debtor countries’ deficits.
- Emphasis on financing patterns:
  - External deficits of debtor countries were financed by capital inflows from within the euro area, particularly from France and Germany.
  - Non-euro-area investors purchased primarily financial instruments issued by other euro-area countries (notably French and German debt securities).
  - Intra-euro-area capital flows funded government debt (Greece), financial sector borrowing (Spain, Ireland), or both (Portugal, Italy).

### Private-sector balance sheets and public debt link
- Rise in current account deficits and external liabilities between euro inception and the global crisis primarily reflected worsening private-sector balance sheets; household net debt rose significantly.
- Financing of private deterioration occurred via:
  - Foreign purchases of domestic government debt previously held by domestic private sector.
  - Increased recourse of debtor countries’ banks to external finance.
- Result: foreign ownership of government debt increased substantially (notably in Greece) even without an increase in many debtor countries’ government debt percent of GDP during 2000–2008.

### Implications for adjustment mechanisms
- Combination of external terms of trade shocks and financing patterns meant key adjustment mechanisms were not operating post monetary union:
  - Trade shocks required REER depreciations in debtor countries; intra-euro-area capital inflows plus euro nominal appreciation contributed to further real appreciation.
  - This real appreciation worsened export performance and delayed necessary external adjustment.

### III — Euro area imbalances and the rest of the world: stylized facts (summary)
- The euro area as a whole remained broadly balanced on the current account but is very open with sizable trade and financial flows vis-à-vis rest of world.
- Investors outside the euro area held portfolio debt claims primarily in core euro-area countries rather than in deficit countries on the eve of the global financial crisis.
- Key empirical observations at EMU accession and during first decade:
  - At euro accession:
    - Greece and Portugal: large current account deficits and REER above historical average.
    - Spain: moderate current account deficit.
    - Italy and Ireland: balanced current account; REER close to historical average.
  - During first decade of EMU:
    - Current account balances worsened significantly in Greece, Ireland, Italy, and Spain; Portugal’s deficit remained very high.
    - Net external liabilities rose sharply, reaching close to or above 100 percent of GDP by end-2010 in Greece, Ireland, Portugal, and Spain.
    - Germany and several Northern European countries built large current account surpluses; euro area overall remained broadly balanced.
  - Growth, saving, investment patterns:
    - Ireland and Spain: high investment rates from construction booms; growth above euro-area average.
    - Greece: stronger growth than rest of euro area; deficit driven by large decline in saving.
    - Portugal: modest growth; declines in investment and household saving.
    - Italy: relatively weak growth; some decline in saving; current account deficit more contained.
  - REER and relative prices:
    - Significant REER appreciations in Greece, Ireland, Portugal, and Spain; smaller in Italy.
    - Consumer prices and unit labor costs rose significantly in periphery relative to core, particularly vis-à-vis Germany.
  - Financial integration and bond yield convergence:
    - Convergence in 10-year government bond yields toward German bunds from mid-1990s until global financial crisis onset (except Greece).

### III.A Real exchange rate appreciation — decomposition and implications
- Decomposition identifies three drivers of RER:
  i. Nominal exchange rate (euro) appreciation.
  ii. Relative price of tradable goods (terms-of-trade shocks).
  iii. Relative wages (domestic wages vs. foreign wages).
- Empirical decomposition findings:
  - Real appreciation in deficit countries primarily reflected strengthening of the euro in all five deficit countries.
  - Spain and Ireland: domestic consumer prices or relative ULC contributed more to REER appreciation than in Greece, Italy, Portugal.
  - France: REER appreciation reflected exclusively nominal effective exchange rate appreciation.
  - Germany: REER remained stable as nominal appreciation was offset by a decline in unit labor costs relative to trading partners.

### III.B Trade developments with non-euro area countries — patterns and shocks
- Broad patterns:
  - Trade with rest of world accounts for large share of imports and exports for debtor countries and for Germany and France.
  - Greece, Italy, Spain: goods trade balance worsened; associated with rapid increase in imports from non-euro-area countries.
  - Ireland: trade surplus grew slower than GDP, particularly vis-à-vis euro area countries.
  - Portugal: stable trade deficit/GDP ratio but rising deficit vis-à-vis non-euro area countries due to terms-of-trade deterioration.
  - Germany: significant trade surplus mainly vis-à-vis non-euro-area countries driven by rapid export increase.
- Sectoral/regional shifts:
  - Germany’s exports to emerging Asia, oil and commodity exporters, and Central and Eastern Europe increased dramatically; exports doubled as ratio to German GDP over 8 years.
  - Italy: export increases but faster import growth led to worsening trade balance.
- Terms of trade and commodity prices:
  - Italy, Portugal, Greece experienced very significant declines in terms of trade during 1999–2008.
  - Crude oil real price rose over 400 percent between 1999 and 2008 (large component of ToT shock).
  - Export prices relative to ULC: until 2006 ULC grew as fast as export prices except in Germany and Spain; after 2006 Italy and Ireland had ULC rising faster than export prices.
  - Note: Greece’s export price to ULC ratio likely biased upward by inclusion of refined oil products.

### III.C Capital flows — financial integration with rest of world
- Euro-area gross external assets and liabilities rose substantially as percent of euro-area GDP between early 2000s and global financial crisis; net position remained fairly stable.
- Despite larger deficits of euro-area debtor countries vis-à-vis rest of world, investors outside euro area concentrated portfolio debt claims in core euro-area countries rather than deficit countries on eve of crisis.
- Implication: financing of euro-area imbalances requires investigation beyond the conjecture that external (non-euro) investors directly financed debtor countries.

### IV — Econometric evidence: hypotheses and empirical strategy
- Hypothesized asymmetric external shocks on euro-area exports:
  1. Rise of China and emerging Asia: displacement of southern European exports in some foreign markets; boosted Germany via machinery/equipment demand.
  2. Higher oil prices: worsened oil trade balances but increased demand from oil producers for machinery (benefiting Germany).
  3. Integration of Central and Eastern Europe: benefited exporters like Germany; increased imports for Southern European countries.
- Data and patterns:
  - Sectoral classification aggregated into High Technology, Medium-High, Medium-Low, Low Technology, and Non-Classified.
  - Germany and Italy had significant specialization in medium-high tech goods at start of period.
  - Germany’s export increase largely from Medium-High Technology goods; almost 1/3 of this increase reflected exports to emerging Asia, oil exporters, and CEE countries.
  - Among sample countries, only Italy had strong medium-high tech export performance besides Germany.

### IV.A Export regression specifications and interpretation
- Aggregate bilateral export regression (preserved form):
  - log(Exports_ijt) = α + β·log(RER_ijt) + δ·log(DomDemand_jt) + F_ij + T_t + ε_ijt
  - RER replaced in robustness checks by bilateral nominal exchange rate and relative CPI.
  - Controls: trade specialization (share of low-technology goods) interacted with trading partner domestic demand; country-specific time trends in robustness checks.
- Coefficients of interest:
  - δ: elasticity of exports w.r.t. trading partner domestic demand.
  - β: price or exchange rate elasticity (real or nominal).
- Sectoral bilateral regressions (preserved form):
  - log(Exports_ijkt) = α + β·log(RER_ijkt) + δ_k·log(DomDemand_jt) + φ·log(DomDemand_it) + F_ij + T_kt + ε_ijkt

### IV.B Regression results — magnitudes and asymmetries (selected quantitative findings)
- Exchange rate elasticities:
  - The 36 percent appreciation of the euro relative to the U.S. dollar from end-1999 to mid-2008 implied a 12–15 percent decrease in exports of euro-area countries to the U.S. on average, and a 20–25 decrease in exports of the debtor countries on average.
- Nominal effective exchange rate appreciation impact (2000–10) using country-specific coefficients:
  - Cumulative loss of export levels of Greece, Italy, Portugal and Spain (in difference from euro-area average effect) was respectively 2.7 percent, 2.8 percent, 1.5 percent and 1.7 percent.
  - Total loss of exports (adding euro-area average effect) from nominal effective exchange rate appreciation was respectively 7 percent, 7.3 percent, 4 percent and 4.6 percent for Greece, Italy, Portugal and Spain.
  - Note: During 2000–10, nominal effective exchange rate appreciation was 13 percent, 13 percent, 7 percent and 8 percent respectively for Greece, Italy, Portugal and Spain.
- Demand elasticity:
  - A 1 percent increase in trading partners’ domestic demand is associated with a 1¼ percent increase in exports on average (Partner Demand ≈ 1.25).
- Heterogeneous demand elasticities:
  - Exports to CEE countries: demand elasticities ≈ 0.6 higher than average.
  - Demand elasticities significantly smaller than euro-area average for Greece and Italy; significantly larger for Spain and Portugal.
  - Exports to China: average export demand elasticity from China significantly higher (by over 0.3) than average elasticity.
  - Exports to oil exporters: export demand elasticities for Greece, Italy, Portugal and Spain are significantly below euro-area average, after accounting for oil-exporter-specific elasticities (which are ≈ 0.2 higher than all trading partners).

### Appendix: sectoral export regressions and displacement (main results)
- Sectoral export regressions (Table A5) key patterns:
  - Export demand elasticities higher for High Tech and Medium-High Tech sectors across trading-partner regions.
  - Example sectoral elasticities for China:
    - Medium-High Tech goods elasticities ≈ 0.05 higher than average.
    - Low Tech goods elasticities ≈ 0.09 below average.
  - Germany: Medium-High and Medium-Low tech goods elasticities ≈ 0.17 above sectoral average for exports to China.
  - Country-specific: Greece sectoral elasticities (excluding “un-classified”) all significantly below average for CEE, China, oil exporters.
  - Italy: sectoral elasticities below average for CEE or oil exporters, but above average for Medium-High Tech goods to China.
- Import (displacement) regressions — testing Chinese displacement:
  - Augmented regression (preserved notation):
    - log(Import_{ijt}) = f(Import_{it}^China, DomDemand_{t}, RER_{ijt}) + controls + errors
  - Interpretation: coefficient μ on Imports_i^China_t:
    - Positive μ → imports from country j positively correlated with imports from China.
    - Negative μ → displacement (Chinese imports substitute for other trading partners).
- Displacement regression results — aggregate and heterogeneous findings:
  - Aggregate-sample: bilateral imports vis-à-vis any trading partner are positively correlated with bilateral imports from China after controlling for partner demand.
    - Magnitudes: full sample — a 10 percent rise in bilateral imports associated with a rise in imports from China of about 2–2½ percent.
    - Euro-area countries only: estimated elasticity ≈ 1.1.
  - Heterogeneous coefficients:
    - Exports of euro-area countries to common markets are negatively correlated with Chinese exports to these markets — suggesting euro-area exports more likely than average to be displaced by Chinese exports.
    - Coefficient smaller for debtor countries — debtor-country exports even more likely to be displaced by Chinese exports.
  - Exchange rate elasticities confirmed: exchange rate elasticity of imports from euro-area debtor countries is larger than average — implying large decline in imports from these countries following effective appreciation.
- Sectoral displacement regressions:
  - Results broadly robust to sectoral specification.
  - For Spain: sectoral data show a negative and large displacement effect larger than aggregate estimate; other debtor countries show smaller sectoral displacement effects.

### Quantification: asymmetric impact of world trade shocks (method and reported magnitudes)
- Method:
  - Average differential elasticity for trading partner j: elast_share_j = sum_k (share_k × elast_{j,k}).
  - Year-t loss of exports to trading partner j (percent of GDP) preserved formula from source:
    - %_export_loss_{j,t} = (1999_export/GDP_1999) × elast_increase_j × (domestic_demand_t / GDP_t) × ((GDP_t / GDP_1999) - 1)
  - Cumulative loss (2000–2008) = sum of annual losses (formula preserved).
- Reported total differential effects on cumulative trade balance, 1999–2008 (in percent of GDP):
  - Aggregate Data:
    - Export demand: Greece -6.8, Italy -7.0, Portugal -1.5, Spain 0.0
    - Displacement of exports: Greece -8.8, Italy -17.1, Portugal -29.2, Spain 2.4
    - Total: Greece -15.6, Italy -24.1, Portugal -30.7, Spain 2.4
  - Sectoral Data:
    - Export demand: Greece -1.3, Italy 0.0, Portugal 0.0, Spain -0.5
    - Displacement of exports: Greece -10.2, Italy -7.4, Portugal 5.9, Spain -25.2
    - Total: Greece -11.5, Italy -7.4, Portugal 5.9, Spain -25.7

### Financing of debtor countries — net foreign asset (NFA) positions and composition
- Main patterns:
  - Net liabilities vis-à-vis other euro-area countries account for lion’s share of increased net external financing for debtor countries since early 2000s.
  - Net liabilities vis-à-vis the United Kingdom account for meaningful part of net liabilities vis-à-vis rest of world (reflects cross-border financial sector activity of core euro-area bank affiliates domiciled in U.K.).
  - Core euro-area countries accumulated net foreign assets within euro area almost exclusively vis-à-vis the five debtor countries.
  - France and Germany: core countries’ claims on debtor countries were mostly in form of debt securities.
- Sectoral/instrument composition of inflows:
  - Purchases of government bonds (particularly in Greece and Portugal).
  - Purchases of bank bonds and lending to domestic banks (particularly Spain, Portugal, Ireland).
  - Italy: largest accumulation of assets overseas reflecting capital outflows by non-bank private sector.
- Net position drivers:
  - General government and financial sector net positions account for lion’s share of increased external liabilities for debtor countries.
  - Worsening household balance sheets associated with increased purchases of nonfinancial assets (primarily housing) and reduced holdings of domestic government debt.
  - Net effect: change more in ownership pattern of domestic public debt (shift to nonresidents) than in overall government debt size; private-sector deterioration drove increased external imbalances.

### Net financial assets by sector (selected table highlights, in percent of GDP, 2001–09 where presented)
- Greece (selected entries): Households 131 59 -71; Government -93 -87 6; Total -42 -102 -60.
- Ireland (selected entries): Households 103 65 -38; Government -13 -28 -15; Total -15 -67 -52.
- Italy (selected entries): Households 20 2186 -16; Government -96 -103 -7; Total 9 -16 -25.
- Portugal (selected entries): Households 140 127 -13; Government -30 -57 -27; Total -48 -106 -58.
- Spain (selected entries): Households 107 76 -31; Government -42 -34 7; Total -34 -90 -56.
- France (selected entries): Households 118 131 14; Government -37 -51 -14; Total 15 -2 -17.
- Germany (selected entries): Households 98 130 32; Government -36 -48 -12; Total 43 0 26.
- Source: Eurostat statistics, OECD statistics (as reported in chapter tables).

### Key regression and graphical findings (selected quantitative coefficients preserved)
- Export regressions (Table 6) core results:
  - Nominal Exchange Rate coefficients reported (examples): 0.404***, 0.304***, 0.343***, 0.204***, 0.348***, 0.406***, 0.310***, 0.412***, 0.385***, 0.325***, 0.404*** (p-values (0.000)).
  - Relative CPI coefficients (examples): -0.437***, -0.405***, -0.378***, -0.294***, -0.381***, -0.442***, -0.414***, -0.447***, -0.428***, -0.408***, -0.441*** (p-values (0.000)).
  - Partner Demand coefficients (examples): 1.294***, 1.296***, 1.268***, 1.261***, 1.209***, 1.203***, 1.219***, 1.228***, 1.220***, 1.232***, 1.295*** (p-values (0.000)).
  - PartnerDemand X LowTechShare: -0.0132* (p-value (0.028)) and related estimates.
  - Country-pair marginal demand interactions include PartnerDemand X R(R=Grc) X P: -1.572*** (p-value (0.000)).
- Displacement (imports from China) regressions (Table 7) core results:
  - Import from China: 0.241***, 0.240***, 0.239***, 0.214***, 0.137***, 0.134***, 0.116*** (p-values (0.000)).
  - Import from China X P(P=Euro): -0.0512*** (p-value (0.000)).
  - Import from China X P(P=Grc): -0.129***, -0.151***, -0.126***, -0.148*** (p-values (0.000)).
  - Import from China X P(P=Esp): 0.0759***, 0.0535*, 0.107***, 0.103*** (p-values (0.001), (0.017), (0.000), (0.000)).
- Sectoral regressions (Table A5, A6) report consistent significance for nominal exchange rate, relative CPI, real exchange rate, partner demand, and interaction terms across sectoral specifications.
- Figures described (figure numbers preserved): Figure 1 (NFA positions 1999–2010), Figure 2 (10-year government bond spreads), Figure 3 (REER decomposition 2000–2010), Figure 4 (Terms of trade and ULC 1999–2008), Figure 5 (International investment position 2001–08), Figures 6–11 (breakdowns of external positions and sectoral NFA positions).

### Policy implications and concluding remarks
- Drivers of divergence:
  - Asymmetric trade developments (rise of China, integration of CEE, rising oil prices) contributed to divergence; southern European exports were negatively affected by Chinese competition.
  - Sharp nominal appreciation of the euro compounded competitiveness losses of deficit countries.
- Financing patterns and implications:
  - Current account deficits of debtor countries were mostly financed by euro-area surplus countries despite trade imbalances vis-à-vis rest of world.
  - Possible reasons for outside investors’ preferences include bailout expectations perceived by core investors; differential access to ECB collateral or financing; index-weight effects for global bond indices.
- Policy recommendations highlighted:
  - Domestic demand-side: fiscal consolidation and internal devaluation.
  - Supply-side: product and labor market reforms to boost productivity and export competitiveness.
  - Need for centralized risk sharing and transfers across euro-area countries to ease adjustment to country-specific shocks; fiscal transfers conditional on strong governance recommended given limited labor mobility and labor market rigidities.
  - External factors that would ease adjustment: stronger external demand, less onerous financing conditions, and depreciation of the euro.
- Overarching conclusion:
  - Build-up in external liabilities reflected multiple factors including external shocks; ease of external financing from core euro-area countries allowed imbalances to persist.
  - Given persistence of external shocks and shift in nonresidents’ willingness to finance large current account deficits, external adjustment in debtor countries is particularly pressing.

*Italic source attribution: Content unit: _wp12236 - References (excerpt) from the IMF PDF _wp12236.pdf*

### References .............................................................................................................

### _wp12236 - References

### Overview
- After its first decade, the euro area faced severe pressure triggered by fiscal adjustments in Greece that spread to Ireland and Portugal and later threatened the eurozone’s existence as Italy and Spain experienced funding pressures.
- The crisis reflected intertwined public debt and banking sector fragilities, weak growth prospects, and substantial gross and net external liabilities—for example, net external liabilities of close to 100 percent of GDP in Greece, Ireland, Portugal, and Spain.
- The paper focuses on the external dimension of the euro area crisis and characterizes factors contributing to growing balance of payments imbalances, emphasizing the five largest “net debtors” in the euro area: Greece, Ireland, Italy, Portugal, and Spain.

### Main findings on causes of imbalances
- Large current account imbalances of individual euro area countries reflected:
  - The asymmetric impact of trade shocks originating outside the euro area.
  - Sustained cheap financing from core euro area countries to the largest net debtors.
- Key external drivers:
  - The rise of China generated strong demand for machinery and equipment goods exported by Germany while displacing exports from euro area debtor countries in their foreign markets.
  - The terms of trade shock associated with higher oil prices contributed to rising trade deficits; higher income in oil producing countries generated strong demand for machinery and equipment exported by Germany.
  - German firms’ outward integration by setting up production platforms in emerging Europe boosted competitiveness and exports of emerging Europe to euro area debtor countries.
- Financing and exchange rate dynamics:
  - Continued easy financing (until the crisis) allowed deficit countries to sustain appreciating real effective exchange rates, which were also driven by the nominal appreciation of the euro, delaying necessary adjustment.

### Distinct contributions relative to existing literature
- Emphasis on trade linkages and relative price dynamics between euro area countries and the rest of the world.
  - While relative price movements within the euro area contributed to debtor countries’ real exchange rate appreciations, the lion share of the appreciation between 2000 and 2009 was accounted for by the nominal appreciation of the euro vis-à-vis other currencies.
  - Econometric evidence presented indicates this appreciation adversely affected export performance of debtor countries, over and above the average impact on euro area exports.
  - Documentation of bilateral trade contributions (China, oil exporters, emerging Europe) to Germany’s growing surplus and debtor countries’ growing deficits.
  - Sectoral export regressions control for all unobserved home country factors that may have affected export performance.
- Emphasis on financing patterns:
  - External deficits of euro area debtor countries (vis-à-vis euro area and non-euro area countries) were financed by capital inflows from within the euro area, in particular from France and Germany.
  - Investors from the rest of the world purchased primarily financial instruments issued by other euro area countries, in particular French and German debt securities.
  - Intra-euro area capital flows financed government debt (in Greece), financial sector borrowing (in Spain or Ireland), or both (in Portugal or Italy).
  - Pattern suggests euro area investors viewed peripheral European securities as closer substitutes for core euro area securities than did investors from outside the euro area.

### Private-sector balance sheets and public debt link
- Rise in current account deficits and external liabilities between the inception of the euro and the global crisis primarily reflected a worsening of private-sector balance sheets, with households’ net debt rising significantly.
- This deterioration was financed directly or indirectly by:
  - Foreign purchases of domestic government debt previously held by the domestic private sector.
  - Increased recourse of debtor countries’ banks to external finance.
- As a result, foreign ownership of government debt increased substantially, particularly in Greece, even though there was no increase in many debtor countries’ government debt in percent of GDP during the period 2000–2008.

### Implications for adjustment mechanisms
- Observed combinations of external terms of trade shocks and financing patterns suggest that, following monetary union, key adjustment mechanisms of debtor countries’ external balances were not operating.
- While trade shocks would have required real effective exchange rate depreciations in debtor countries to restore external sustainability in the long run, intra-euro area capital inflows and the trend in the euro nominal exchange rate contributed instead to further real appreciation, which further affected export performance.

_Italic source attribution: Content unit: _wp12236 - References (excerpt) from the IMF PDF _wp12236.pdf_

### Section III presents an ancillary set of stylized facts which emphasize trade and financial

### _wp12236 - Section III presents an ancillary set of stylized facts which emphasize trade and financial

### III. EURO AREA IMBALANCES AND THE REST OF THE WORLD: NEW STYLIZED FACTS — Summary of major findings
- Trade and financial linkages between the euro area and the rest of the world played an important role in explaining external imbalances of individual euro area countries.
- The euro area as a whole remained broadly balanced on the current account while being a very open economy with sizable trade and financial flows vis-à-vis the rest of the world.
- Investors from outside the euro area held their portfolio debt claims primarily in “core” euro area countries rather than in deficit countries on the eve of the global financial crisis (see Figure 6).

### II. STYLIZED FACTS ON EURO AREA IMBALANCES — Key empirical observations
- At euro accession:
  - Greece and Portugal’s current account deficits were already large and had a real effective exchange rate (REER) above historical average.
  - Spain had a moderate current account deficit.
  - Italy and Ireland had a balanced current account; their REER were close to historical average.
- During the first decade of EMU:
  - Current account balances in Greece, Ireland, Italy, and Spain worsened significantly; Portugal’s deficit remained at very high levels (Table 2).
  - Net external liabilities rose sharply, reaching levels close to or above 100 percent of GDP by end-2010 in Greece, Ireland, Portugal, and Spain (Figure 1).
  - Germany and several Northern European countries built large current account surpluses; euro area as a whole remained broadly balanced.
- Growth, saving, and investment patterns:
  - Ireland and Spain: investment rates boosted by construction booms; growth considerably above euro area average, aided by rising labor forces.
  - Greece: stronger growth than rest of euro area; widening current account deficit mostly due to a large decline in saving.
  - Portugal: modest growth with declines in both investment and household saving.
  - Italy: relatively weak growth, some decline in saving; current account deficit in percent of GDP more contained than in other countries.
- Real effective exchange rate (REER) movements and relative prices:
  - Significant REER appreciations in Greece, Ireland, Portugal, and Spain; to a lesser extent in Italy (Figure 3).
  - Consumer prices and unit labor costs rose very significantly in the euro area periphery relative to the euro area core, particularly vis-à-vis Germany (Table A4).
- Financial integration and bond yield convergence:
  - Convergence in 10-year government bond yields toward German bunds began in the mid-1990s and persisted until the onset of the global financial crisis (Figure 2). With the exception of Greece, most reduction of bond spreads in Southern Europe occurred in run-up to EMU.

### B. Traditional explanations for imbalances — Mechanisms and inconsistencies
- Two complementary traditional explanations:
  1. Financial integration and expectations of convergence within the euro area (capital inflows from richer to poorer countries).
  2. “Over-optimism” and wage/price rigidities in borrowing countries leading to strong domestic demand and faster rises in domestic prices and labor costs (intra-euro area competitiveness losses).
- Neoclassical convergence implications:
  - Removing transaction costs and currency risk should lead to net capital inflows to less advanced countries, associated with rising domestic investment and/or declining savings.
  - Real appreciation can follow from higher consumption of non-tradable goods as productivity or income rises (Balassa-Samuelson).
- Empirical tensions with pure neoclassical story:
  - Greece and Portugal experienced declines in corporate saving concurrent with declines in domestic investment—difficult to reconcile with higher marginal product of capital.
  - Germany’s rising surplus mainly reflected a rise in corporate savings and a decline in domestic investment.
- Over-optimism and real appreciation:
  - Financial liberalization, lower credit constraints, and over-optimistic expectations contributed to domestic price and ULC increases inconsistent with productivity gains.
  - Asset-financed real estate booms sustained growth of non-tradable sectors in some countries.
  - Jaumotte and Sodsriwiboon (2010) find deficits in excess of fitted values related to a "euro effect" and to 1990s financial reforms.
  - Resulting REER appreciation contributed to crowding out manufacturing and exports (Portugal: decade of low productivity gains and stagnant growth; Greece and Spain: growth sustained by strong domestic demand and large current account deterioration).
  - Bilateral trade imbalances among euro area countries became more persistent, especially in countries with more rigid labor markets (Berger and Nitsch, 2010).

### III.A Real exchange rate appreciation — Decomposition and implications
- Decomposition of CPI-based REER in a 3-country environment (home, rest of euro area, rest of world):
  - RER depends on: domestic price level (P), euro nominal exchange rate (S), price levels in euro area (EA P*) and non-euro area trading partners (NEA P*), and trade share with non-euro area countries (α).
  - Rearranged expression shows RER affected by:
    i. Nominal exchange rate: nominal appreciation tends to appreciate RER by improving terms of trade vis-à-vis non-euro area countries but worsens relative wage competitiveness.
    ii. Relative price of tradable goods: an increase in price of foreign imported goods (negative terms of trade shock) tends to depreciate RER.
    iii. Relative wages: increase in domestic wages relative to foreign wages tends to appreciate RER.
- Empirical decomposition finding:
  - Real appreciation in deficit countries primarily reflected the strengthening of the euro in all five current account deficit countries.
  - Spain and Ireland: domestic consumer prices (or relative ULC) contributed more significantly to REER appreciation than in Greece, Italy, and Portugal.
  - France: REER appreciation reflected exclusively nominal effective exchange rate appreciation.
  - Germany: REER remained stable; nominal appreciation was offset by a decline in unit labor costs relative to trading partners.
- Interpretation:
  - Part of the euro appreciation and ensuing loss in competitiveness for periphery countries can be viewed as an “external shock” given that non-euro area investors invested mainly in core countries (see Section V).

### III.B Trade developments with non-euro area countries — Patterns and shocks
- Broad patterns (Table 3 and Table 4):
  - Trade with rest of world accounts for a large share of imports and exports for debtor countries and for Germany and France.
  - Greece, Italy, Spain: goods trade balance worsened during the decade; worsening associated with rapid increase in imports from non-euro area countries.
  - Ireland: trade surplus grew at a slower rate than GDP, particularly vis-à-vis euro area countries.
  - Portugal: stable trade deficit to GDP ratio but rising trade deficit vis-à-vis non-euro area countries (mainly terms-of-trade deterioration); export-to-GDP ratios vis-à-vis euro and non-euro area countries stable or declining.
  - Germany: significant trade surplus mainly vis-à-vis non-euro area countries driven by rapid increase in exports.
- Sectoral and regional shifts (Table 4):
  - Dramatic increase in Germany’s exports to emerging Asia, oil and commodity exporters, and Central and Eastern Europe; exports doubled as a ratio of German GDP over 8 years.
  - German trade balance vis-à-vis commodity exporters improved despite a dramatic increase in commodity prices.
  - Greece, Spain, France: modest export increases to these regions and substantial deterioration in trade balances.
  - Italy: export increases but faster import growth led to worsening trade balance.
  - Portugal: similar increases of exports and imports as percent of GDP.
- Terms of trade and commodity prices:
  - Italy, Portugal, and Greece experienced very significant declines in their terms of trade during 1999–2008 (Figure 4).
  - A large component of the terms of trade shock came from the steady and substantial increase of real crude oil prices (over 400 percent between 1999 and 2008).
  - Export prices relative to unit labor costs: until 2006, ULC grew as fast as export prices except in Germany and Spain; after 2006, Italy and Ireland had ULC rising faster than export prices.
  - Note: Greece’s ratio of export price to ULC likely biased upwards by inclusion of refined oil products in export price index.

### III.C Capital flows — Financial integration with the rest of the world
- Euro area gross external assets and liabilities rose substantially in percent of euro area GDP between the beginning of the decade and the onset of the global financial crisis, while net position remained fairly stable (Figure 5).
- Despite larger deficits of euro area debtor countries vis-à-vis the rest of the world, investors from outside the euro area concentrated portfolio debt claims in core euro area countries rather than deficit countries on the eve of the global financial crisis (Figure 6).
- Implication: financing of euro area imbalances warrants in-depth investigation beyond simplistic conjectures that external investors directly financed debtor countries.

### IV. ECONOMETRIC EVIDENCE — Hypothesis and empirical strategy
- Hypothesized external shocks with asymmetric effects on euro area exports:
  1. Rise of China and emerging Asia: may have displaced southern European exports from some foreign markets while boosting demand for machinery and equipment (benefiting Germany).
  2. Higher oil prices: worsened oil trade balances for all countries but fast income growth in commodity exporters may have benefited countries such as Germany exporting goods in high demand by oil producers.
  3. Integration of Central and Eastern Europe into euro area production chains: may have benefited exporters like Germany (through FDI and lower wages) while increasing imports for Southern European countries.
- Data and descriptive patterns:
  - Table 5 (5-digit sectoral classification aggregated into High Technology, Medium-High, Medium-Low, Low Technology, and Non-Classified) shows Germany and Italy had significant specialization in medium-high tech goods at the start of the period.
  - Germany’s strong export performance largely accounted for by rapid increase in Medium-High Technology goods, with almost 1/3 of such increase reflecting exports to emerging Asia, oil exporters, and CEE countries.
  - Among sample countries, only Italy had strong performance in medium-high tech exports besides Germany.

### IV.A Econometric analysis — Export regressions: specification and interpretation
- Aggregate bilateral export regression specification:
  - log(Exports_ijt) = α + β·log(RER_ijt) + δ·log(DomDemand_jt) + F_ij + T_t + ε_ijt
  - RER replaced in robustness checks by bilateral nominal exchange rate and relative CPI to separate nominal exchange rate effects.
  - Controls: trade specialization (share of low-technology goods) interacted with trading partner domestic demand; country-specific time trends in robustness checks.
- Coefficients of interest:
  - δ: elasticity of exports with respect to trading partner domestic demand.
  - β: price or exchange rate elasticity (real or nominal).
- Sectoral bilateral regressions:
  - log(Exports_ijkt) = α + β·log(RER_ijkt) + δ_k·log(DomDemand_jt) + φ·log(DomDemand_it) + F_ij + T_kt + ε_ijkt
  - δ_k: sector-specific elasticity of foreign demand; sectoral time dummies control for sector trends.

### IV.B Regression results — Magnitudes and asymmetries
- Price/exchange rate elasticities:
  - Estimated elasticities imply the 36 percent appreciation of the euro relative to the U.S. dollar from end-1999 to mid-2008 implied a 12–15 percent decrease in exports of euro area countries to the U.S. on average, and a 20–25 decrease in exports of the debtor countries on average.
- Nominal effective exchange rate appreciation impact (2000–10):
  - Using country-specific coefficients (column (4)), cumulative loss of export levels of Greece, Italy, Portugal and Spain (in difference from a euro area average effect) was respectively 2.7 percent, 2.8 percent, 1.5 percent and 1.7 percent.
  - Total loss of exports (adding the euro area average effect) from nominal effective exchange rate appreciation was respectively 7 percent, 7.3 percent, 4 percent and 4.6 percent for Greece, Italy, Portugal and Spain.
  - Note: During 2000-10, the nominal effective exchange rate appreciation was 13 percent, 13 percent, 7 percent and 8 percent respectively for Greece, Italy, Portugal and Spain.
- Demand elasticity:
  - A one percent increase in trading partners’ domestic demand is associated with a 1¼ percent increase in exports on average.
- Heterogeneous demand elasticities by trading partner region and reporting country (columns 5–12):
  - Exports to CEE countries: demand elasticities significantly larger than average (by about 0.6).
  - Cross-country heterogeneity:
    - Demand elasticities significantly smaller than euro area average for Greece and Italy.
    - Demand elasticities significantly larger for Spain and Portugal.
    - No compelling evidence that Germany benefits from higher export demand elasticities than other euro area countries.
  - Trade specialization effect: countries with higher export shares in low-tech goods tend to have smaller demand elasticities from CEE countries.
  - Exports to China:
    - Average export demand elasticity from China significantly higher (by over 0.3) than average elasticity from all trading partners.
    - Italy’s export demand elasticity vis-à-vis China is significantly lower than the euro area average elasticity.
  - Exports to oil exporters:
    - Export demand elasticities for Greece, Italy, Portugal and Spain are significantly below the euro area average, after accounting for oil-exporter-specific elasticities (which are significantly higher than all trading partners by about 0.2).
- Overall conclusion from regressions:
  - Trade shocks from emerging Asia, oil exporters, and CEE countries had asymmetric impacts across euro area countries.
  - Southern European countries experienced export demand shortfalls relative to the euro area average, while Germany benefitted disproportionately from demand shifts to particular regions and from sectoral composition (medium-high tech specialization).

*Source: _wp12236 - Section III presents an ancillary set of stylized facts which emphasize trade and financial*

### Appendix Table A5 reports export regressions estimated at the sectoral level, allowing us to

### _wp12236 - Appendix Table A5 reports export regressions estimated at the sectoral level, allowing us to

### Export regressions — sectoral specification and main findings
- Regression framework:
  - Controls: reporting countries’ domestic demand, country pair dummies, sectoral time dummies.
  - Interaction terms: trading-partner-specific export demand elasticities (CEE countries, emerging Asia, oil exporters); these elasticities allowed to vary across sectors and across euro area reporting countries (Germany and the debtor countries).
  - Omitted sector: the “un-classified” (e.g., agriculture and mining).
- Key empirical findings:
  - On average, export demand elasticities are higher for High Tech and Medium-High Tech sectors than for other sectors for each trading-partner region (CEE countries, China and oil exporters).
  - Example elasticities reported:
    - For China (columns 5–6): Medium-High Tech goods elasticities are about 0.05 higher than average.
    - For China: Low Tech goods elasticities are about 0.09 below average.
    - Germany: export demand elasticities of Medium-High Tech and Medium-Low Tech goods are about 0.17 above the sectoral average of euro area countries for goods exported to China.
  - Country-specific patterns:
    - Greece: sectoral elasticities (excluding “un-classified”) are all significantly below average for all three trading-partner regions.
    - Italy: sectoral elasticities are below average for trade with CEE countries or oil exporters but above average for exports of Medium-High Tech goods to China.
  - General conclusion: strong evidence that sectoral export demand elasticities are significantly below the sectoral average for many sectors and many euro area debtor countries.

### Import regressions — testing displacement by Chinese imports
- Augmented import regression specification (notation preserved from source):
  - log(Import_{ijt}) = f(Import_{it}^China, DomDemand_{t}, RER_{ijt}) + controls + errors
  - Interpretation: coefficient μ on Imports_i^China_t: positive → imports from country j positively correlated with imports from China; negative → displacement effect (Chinese imports substitute for other trading partners).
  - Additional control: log( low_techImports_{jiChina,t} ) where low_j is the share of low technology goods in exports of country j (to account for trade specialization).
  - Sectoral-level estimations: estimate sector-specific elasticity μ_k of euro area exports to market i relative to Chinese exports to the same market, controlling for partner domestic demand, sector dummies, and country dummies.
- Regression-sample and scope:
  - Sample: 17 major trading partners of euro area countries, 1999–2009.
  - Reported regressions: columns 1–7 (complete sample), columns 8–9 (euro area reporting countries).

### Displacement regression results and interpretations
- Aggregate-sample results:
  - On average, bilateral imports vis-à-vis any trading partner are positively correlated with bilateral imports from China, even after controlling for total domestic demand of the trading partner.
  - Magnitudes reported:
    - Full sample: a 10 percent rise in bilateral imports is associated with a rise in imports from China of about 2–2½ percent.
    - Euro area countries only: estimated elasticity is 1.1 (much smaller).
  - Interpretation: positive coefficient inconsistent with average displacement; imports from China may proxy for demand effects not captured by total domestic demand.
- Heterogeneous coefficients and country-specific findings:
  - Allow μ to vary across trading partners and reporting countries; control for partner trade specialization (low-tech shares).
  - Findings for euro area reporting countries:
    - Exports of euro area countries to common markets are negatively correlated with Chinese exports to these markets (columns 4–6) — suggesting euro area exports are more likely than average to be displaced by Chinese exports.
    - Coefficient smaller for debtor countries (columns 4, 5, and 8) — debtor-country exports even more likely to be displaced by Chinese exports.
  - Trade specialization (low-tech share) generally shows no significant displacement effects, except in regressions focusing only on euro area reporting countries (limited coverage for non-euro area countries).
  - Exchange rate elasticities:
    - Confirmed (columns (2) and (10)) that the exchange rate elasticity of imports from euro area debtor countries is larger than the average elasticity — implying a large decline in imports from these countries following effective appreciation of their nominal exchange rate.

### Sectoral displacement regressions (Appendix Table A6) and robustness
- Sectoral regressions:
  - Controls: trading partner’s domestic demand and sector-specific time dummies.
  - Main findings remain broadly robust to sectoral specification.
- Notable sectoral result:
  - For Spain: negative and large displacement effect obtained with sectoral data (larger than with aggregate data), while other three debtor countries show smaller displacement effects in sectoral regressions.
  - Explanation note: estimated displacement effects may be smaller with sectoral data if part of the displacement effect reflects pure sectoral patterns absorbed by sector time dummies; for Spain, aggregation bias may have attenuated displacement effects in aggregate regressions.

### Quantification: asymmetric impact of world trade shocks
- Method to compute export losses:
  - Average differential elasticity for trading partner j: elast_share_j = sum_k (share_k × elast_{j,k}), where share_k is sector k share in total exports and elast_{j,k} is the differential elasticity of sector k (relative to euro area average) for exports to region j.
  - Year-t loss of exports to trading partner j (as percent of GDP):
    - %_export_loss_{j,t} = (1999_export/GDP_1999) × elast_increase_j × (domestic_demand_t / GDP_t) × ((GDP_t / GDP_1999) - 1)
    - (formula preserved as in source text)
  - Cumulative loss in percent of GDP between 2000 and 2008 = sum of annual losses.
- Aggregate magnitude statement from source:
  - “Total cumulative differential effects of trade developments vis-à-vis China, emerging Europe and oil exporters on the trade balance of euro area debtor countries appear to be quite large,” with variation across regression specifications.
  - Sectoral-data estimates: export demand and displacement effects generally smaller, except displacement effect larger for Spain.

### Total differential effects on the cumulative trade balance, 1999–2008 (in percent of GDP) — reported table values
- Differential Effect :GreeceItalyPortugal   Spain
  - Aggregate Data:
    - Export demand: -6.8 -7.0 -1.5 0.0
    - Displacement of exports: -8.8 -17.1 -29.2 2.4
    - Total: -15.6 -24.1 -30.7 2.4
  - Sectoral Data:
    - Export demand: -1.3 0.0 0.0 -0.5
    - Displacement of exports: -10.2 -7.4 5.9 -25.2
    - Total: -11.5 -7.4 5.9 -25.7

### Financing of euro area debtor countries — net foreign asset (NFA) positions and composition
- Main patterns:
  - Net liabilities vis-à-vis other euro area countries account for the lion’s share of increased net external financing for euro area debtor countries since early 2000s.
  - Net liabilities vis-à-vis the United Kingdom account for a meaningful part of net liabilities vis-à-vis the rest of the world — partly reflecting cross-border financial sector activity of affiliates of core euro area banks domiciled in the U.K.
  - Mirror image: core euro area countries accumulated net foreign assets vis-à-vis debtor countries and net foreign liabilities vis-à-vis the rest of the world.
  - France and Germany: accumulated net foreign assets within the euro area almost exclusively vis-à-vis the five debtor countries; Germany’s net asset position vis-à-vis the rest of the world grew positive, while France had a growing net liability vis-à-vis the rest of the world.
  - Core countries’ claims on debtor countries were mostly in the form of debt securities.
- Sectoral and instrument composition:
  - Capital inflows destination:
    - Purchases of government bonds (particularly in Greece and Portugal).
    - Purchases of bank bonds and lending to domestic banks (particularly Spain, Portugal, Ireland).
    - Italy: largest accumulation of assets overseas, reflecting capital outflows by the non-bank private sector.
  - Net position drivers: general government and financial sector net positions account for the lion’s share of increased external liabilities for debtor countries.
  - Domestic private-sector balance-sheet deterioration:
    - Worsening external position largely associated with worsening household financial balance sheets, mostly explained by increased purchases of nonfinancial assets (primarily housing).
    - Private sector substantially reduced holdings of domestic government debt and increased indebtedness vis-à-vis domestic financial system — increasing reliance on external funding.
  - Net effect: the change was more in ownership pattern of domestic public debt (shift to nonresidents) than in overall size; worsening private-sector balance sheets drove increased external imbalances.

### Policy implications and concluding remarks
- Drivers of divergence in external balances:
  - Asymmetric impact of trade developments: rise of China, integration of Central and Eastern Europe, rising oil prices — contributed to divergence.
  - Exports of several Southern European countries were negatively affected by Chinese competition; Chinese import demand provided little benefits to their trade balances.
  - Sharp nominal appreciation of the euro compounded competitiveness losses of deficit countries within the euro area.
- Financing patterns and implications:
  - Current account deficits of debtor countries were mostly financed by euro area surplus countries despite trade imbalances vis-à-vis the rest of the world — highlighting intra-euro area financial integration’s special role.
  - Possible explanations for outside investors’ portfolio preferences: bailout expectations perceived by core euro area investors; differential access to ECB collateral or financing; index-weight effects for global bond indices.
- Policy recommendations highlighted:
  - Domestic demand-side policies: fiscal consolidation and internal devaluation.
  - Supply-side policies: product and labor market reforms to boost productivity and export competitiveness.
  - Need for centralized risk sharing and transfers across euro area countries to ease adjustment to country-specific shocks; fiscal transfers conditional on strong governance would be particularly important given limited labor mobility and labor market rigidities.
  - External factors that would ease adjustment: stronger external demand, less onerous financing conditions, and depreciation of the euro.
- Overarching conclusion:
  - The build-up in external liabilities in several euro area countries reflected a variety of factors, including external shocks, with the ease of external financing coming from core euro area countries playing an important role in allowing these imbalances to persist.
  - Given persistence of external shocks and the shift in nonresidents’ willingness to finance large current account deficits, the need for external adjustment in debtor countries is particularly pressing.

*Source: IMF staff analysis and regression results drawn from the content unit provided.*

### 1. Emerg in g  A s ia0.0%0.0%0.0%0.0%0.1%0.1%0.0%0.0%0.0%0.0%

### _wp12236 - 1. Emerg in g  A s ia0.0%0.0%0.0%0.0%0.1%0.1%0.0%0.0%0.0%0.0%

### Export regressions — core results (Table 6)
- Dependent variable: annual bilateral export volume of 11 euro countries with their top 50 trading partners from 1990 to 2009.
- Sample: 11 euro countries are Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Netherlands, Portugal, and Spain. Observations: 5802.
- Goodness of fit: R-sq uared reported as 0.981, 0.982 across specifications.
- Robust pval in parentheses; significance coding: *** p<0.01, ** p<0.05, * p<0.1.
- Estimated coefficients (selected, preserved exactly as reported):
  - No minal Exchange Rate: 0.404***, 0.304***, 0.343***, 0.204***, 0.348***, 0.406***, 0.310***, 0.412***, 0.385***, 0.325***, 0.404*** (all p-values reported as (0.000)).
  - Relative CPI: -0.437***, -0.405***, -0.378***, -0.294***, -0.381***, -0.442***, -0.414***, -0.447***, -0.428***, -0.408***, -0.441*** (all p-values (0.000)).
  - Real Exchange Rate: 0.425***, 0.366*** (p-values (0.000)); in an interaction specification Real Exchange Rate: 0.141** (p-value (0.009)).
  - Partner Demand: 1.294***, 1.296***, 1.268***, 1.261***, 1.209***, 1.203***, 1.219***, 1.228***, 1.220***, 1.232***, 1.295***, 1.290***, 1.286*** (all p-values (0.000)).
  - PartnerDemand X LowTechShare (r): -0.0132* (p-value (0.028)); other related estimates shown as -0.00967 (p-value (0.119)), -0.00855 (p-value (0.143)).
  - PartnerDemand X P (various P definitions): 0.594***, 0.632***, 0.603***, 0.372***, 0.370***, 0.356**, 0.169*, 0.113, 0.171** (p-values reported correspondingly).
- Country-pair and region interaction effects (marginal differences in demand elasticity):
  - PartnerDemand X R(R=Grc) X P: -1.572*** (column (7), p-value (0.000)) — interpreted as the marginal difference in demand elasticity for exports from Greece to CEE countries.
  - PartnerDemand X R(R=Its) X P: -0.778***, -0.858***, -0.353* (p-values (0.000), (0.000), (0.016) respectively).
  - PartnerDemand X R(R=Prt) X P: 0.754***, -0.144, -0.657** (p-values (0.000), (0.753), (0.008)).
  - PartnerDemand X R(R=Esp) X P: 0.948***, -0.336, -0.323** (p-values (0.000), (0.159), (0.007)).
- All regressions include a full set of country fixed effects and time fixed effects. Independent variables include nominal exchange rate, relative CPI (reporter CPI/Partner CPI), real exchange rate, partner demand (trading partner's domestic demand), reporter's export shares in low-tech goods, and interactions.

### Displacement effect — imports from China and impacts on euro-area trade (Table 7)
- Dependent variable: annual bilateral import volume of 17 countries with their top 50 trading partners from 1990 to 2009. The 17 countries: United States, United Kingdom, Austria, Belgium, Denmark, France, Germany, Italy, Netherlands, Sweden, Switzerland, Japan, Spain, Turkey, Russian Federation, Czech Republic, and Poland. Observations: typically 3751; sub-samples 2270 for “only euro” partner specifications.
- R-squared: 0.982–0.988 across specifications.
- Key coefficients (preserved exactly):
  - Nominal Exchange Rate: 0.318***, 0.341***, 0.342***, 0.333***, 0.352***, -0.247***, -0.272*** (p-values (0.000) except as reported).
  - Relative CPI: -0.354***, -0.399***, -0.400***, -0.392***, -0.413***, 0.168**, 0.193*** (p-values as reported).
  - Real Exchange Rate: 0.341***, 0.423***, -0.134* (p-values (0.000), (0.000), (0.041)); Real Exchange Rate X GIPS: -0.622***, -0.264** (p-values (0.000), (0.003)).
  - ReporterDemand: 1.803***, 1.763***, 1.853***, 1.436***, 1.437***, 1.436***, 1.458***, 1.125***, 1.119***, 1.044*** (all p-values (0.000)).
  - Import from China: 0.241***, 0.240***, 0.239***, 0.214***, 0.137***, 0.134***, 0.116*** (all p-values (0.000)).
  - Import from China X LowTechShare(p): 0.00227 (p-value (0.717)), 0.00116 (0.853), 0.000948 (0.880), -0.0160* (0.021), -0.0189** (0.006), -0.0183** (0.007).
  - Import from China X P(P=Euro): -0.0512***, -0.0513***, -0.0515*** (p-values (0.000)).
  - Import from China X P(P=GIPS): -0.0418** , -0.0421** , -0.0201 (p-values (0.004), (0.004), (0.143)).
  - Import from China X P(P=Grc): -0.129***, -0.151***, -0.126***, -0.148*** (p-values (0.000)).
  - Import from China X P(P=Ita): -0.0657***, -0.0881***, -0.0462**, -0.0531*** (p-values (0.000), (0.000), (0.002), (0.000)).
  - Import from China X P(P=Prt): -0.151***, -0.173***, -0.116***, -0.109*** (p-values (0.000)).
  - Import from China X P(P=Esp): 0.0759***, 0.0535*, 0.107***, 0.103*** (p-values (0.001), (0.017), (0.000), (0.000)).
- Interpretation note provided in the source: for example, -0.0512 for “Import from China X P(P=Euro)” means the marginal effect of imports from China on the reporter country’s imports from euro countries. Columns (7) and (8) use the sub-sample where trading partners are only euro countries.

### Net financial assets by sector (Table 8) — 2001–09, in percent of GDP (selected country sectoral positions and changes)
- Greece:
  - Households: 131 59 -71
  - Government: -93 -87 6
  - Financial Sector: -9 -64
  - Non-financial Sector: -71 -69 2
  - To tal: -42 -102 -60
- Ireland:
  - Households: 103 65 -38
  - Government: -13 -28 -15
  - Financial Sector: -213
  - Non-financial Sector: -103 -105 -2
  - To tal: -15 -67 -52
- Italy:
  - Households: 20 2186 -16
  - Government: -96 -103 -7
  - Financial Sector: 219 17
  - Non-financial Sector: -99 -117 -18
  - Total: 9 -16 -25
- Portugal:
  - Households: 140 127 -13
  - Government: -30 -57 -27
  - Financial Sector: -10 -19
  - Non-financial Sector: -148 -174 -26
  - To tal: -48 -106 -58
- Spain:
  - Households: 107 76 -31
  - Government: -42 -34 7
  - Financial Sector: 31 17
  - Non-financial Sector: -103 -143 -40
  - To tal: -34 -90 -56
- France:
  - Households: 118 131 14
  - Government: -37 -51 -14
  - Financial Sector: 111 98
  - Non-financial Sector: -77 -102 -25
  - Total: 15 -2 -17
- Germany:
  - Households: 98 130 32
  - Government: -36 -48 -12
  - Financial Sector: 0 77
  - Non-financial Sector: -58 -59 -1
  - To tal: 43 0 26
- Source: Eurostat statistics, OECD statistics.

### Key graphical findings described (figures and trends)
- Figure 1: Net Foreign Asset Positions 1999–2010, in Percent of GDP — series for Greece, Ireland, Italy, Portugal, Spain, France, Germany (source: IFS data).
- Figure 2: Ten-Year Government Bond Spreads Against German Bunds — series for Ireland, Italy, Portugal, Spain, Greece (right); axes showing ranges including ‐5.00 to 40.00 and -2.00 to 18.00 (source: DataInsight).
- Figure 3: Decomposition of Real Effective Exchange Rates, Percentage Change from 2000 to 2010 — ULC-based REER (Eurostat, 36 trading partners) and CPI-based REER (INS); plotted ranges -20% to 30% for NEER and REER components.
- Figure 4: Terms of Trade and Unit Labor Costs — ToT (Goods and Services) index (100=2000) and Crude Oil (RHS) for Spain, Germany, France, Ireland, Italy, Portugal, Greece; index ranges and years 1999–2008 (sources: Eurostat and IMF, World Economic Outlook database).
- Figure 5: The International Investment Position of the Euro Area 2001–08 (In percent of Euro Area GDP) — assets, liabilities, NFA (right axis) with ranges reported from -250% to 250% and NFA axis -20% to 0% (source: Waysand et al. (2010), World Economic Outlook and authors' calculations).
- Figures 6–11: Various breakdowns of external positions and sectoral net foreign asset positions, including:
  - Figure 6: Share of Outstanding Debt Securities Held Outside the Euro Area (2008) (authors’ calculations).
  - Figure 7: Net Foreign Assets of Euro Area Debtor Countries (In percent of Euro Area GDP) 2001–2008.
  - Figure 8: Net Foreign Assets of “Core” Euro Area Countries (In percent of Euro Area GDP) 2001–2008.
  - Figure 9: Bilateral Net Foreign Assets of Germany and France 2001–2008 (percent of each country’s GDP) with components GIIPS, ROW + unallocated, rest of EA excl GIIPS, total.
  - Figure 10: Bilateral Net Foreign Assets of Germany and France vis-à-vis Euro Area Debtor Countries (By instrument) — Other Investment, Portfolio Debt Investment, Portfolio Equity Investment, Direct Investment (bil. US$) 2001–2008.
  - Figure 11: Sectoral Net Foreign Asset Positions (In percent of GDP) series for Greece, Ireland, Italy, Portugal, Spain, France, Germany by sector (Other, Banks, General government, Monetary authorities) for 1999–2010. Note: Data available only from 2001 for Ireland.
- Sources for figures: IFS data; DataInsight; Eurostat; INS; Waysand et al. (2010); World Economic Outlook; authors' calculations.

### Analytical implications and interpretation (from tables and notes)
- Exchange rate and price competitiveness:
  - Nominal exchange rates and relative CPI coefficients consistently significant, indicating exchange rate movements and relative inflation are important determinants of bilateral export and import volumes.
  - Real exchange rate coefficients are positive and significant in several specifications, with negative interactions for GIPS for some specifications (Real Exchange Rate X GIPS: -0.622***, -0.264**), indicating heterogeneity in real exchange rate effects across reporter groups.
- Demand elasticities and country heterogeneity:
  - Partner Demand coefficients are large, positive, and highly significant (around 1.2–1.3 range), underscoring the strong role of partner domestic demand in driving exports.
  - Interaction terms show substantial variation in demand elasticities by reporter and partner pairs (e.g., large negative coefficients for Greece interactions), implying that shocks to partner demand have uneven effects across euro-area exporters.
- China’s role and displacement:
  - Import from China has a positive direct coefficient on reporter imports (e.g., 0.241***), but interaction terms with partner groups show negative marginal effects on imports from euro countries (e.g., Import from China X P(P=Euro): -0.0512***), indicating displacement of euro-area exports by Chinese imports for some partner relationships.
  - Heterogeneous displacement: statistically significant negative interactions for specific partner sets (Grc, Ita, Prt) and positive or mixed signs for Esp in some specifications.
- Sectoral and cross-country external positions:
  - Table 8 and figures document large cross-country variation in net foreign asset and sectoral positions between 2001 and 2009, with notable swings for Greece, Ireland, Portugal, Spain, and differences versus Germany and France.
  - Total net positions changed markedly (examples: Greece Total -42 -102 -60; Ireland Total -15 -67 -52; Germany Total 43 0 26), reflecting balance-sheet transformations over the 2001–09 period.

*Source: IMF staff calculations based on tables and figures in the provided chapter content.*

### References

### _wp12236 - References

### References (selected)
- Bibliographic citations include works by Baldwin; Baumann and di Mauro; Bayoumi; Bordo, Markiewicz and Jonung; Bennett et al.; Berger and Nitsch; Blanchard and Giavazzi; Blanchard; Blank and Buch; Choi; Coeurdacier and Martin; Chinn; De Santis and Gérard; Di Mauro and Forster; Di Mauro, Forster, and Lima; Flam and Nordström; Goldstein and Khan; Giavazzi and Spaventa; International Monetary Fund (2011, "Regional Economic Outlook: Europe,” May); Jappelli and Pagano; Jaumotte and Sodsriwiboon; Kalemli-Ozcan, Papaioannou, and Peydró; Lane; Lane and Milesi-Ferretti; Marquez; Marin; Micco, Stein and Ordoñez; OECD (2005); Rose; Schmitz and von Hagen; Spiegel; Waysand, Ross, and de Guzman; Westerlund.
- Journals and working-paper series cited include ECB Working Papers, IMF Working Papers, NBER Working Papers, Brookings Papers on Economic Activity, Portuguese Economic Journal, Comparative Economic Studies, Journal of International Money and Finance, Journal of the Japanese and International Economies, Open Economies Review, Handbook of International Economics, CEPR Discussion Papers, Journal of International Economics, Review of Economics and Statistics, Economic Policy, Oxford Bulletin of Economics and Statistics, and others.

### Appendix — Data sources, sample, and variable construction
- Bilateral export data: IMF’s Direction of Trade Statistics.
- Export regressions: annual bilateral exports of 11 euro countries (Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Netherlands, Portugal, and Spain) to their top 50 trading partners.
- Displacement effect regressions: focus on countries that choose to import from either the euro area or China. Selected countries are those in the top 20 export trading partners for at least 6 euro countries: United States, United Kingdom, Austria, Belgium, Denmark, France, Germany, Italy, Netherlands, Sweden, Switzerland, Japan, Spain, Turkey, Russia, Czech Republic, and Poland (17 countries). Bilateral nominal exports and imports for those 17 counties with their top 50 trading partners were collected.
- Price deflators: nominal bilateral exports (imports) are converted into real values using the reporter country’s export (import) price deflators because bilateral price deflators are not available.
- Sectoral trade data: U.N. Commodity Trade Statistics database (UN COMTRADE). Product types aggregated into four categories according to technology intensity using the ISIC Rev. 3 breakdown: high, medium-high, medium-low, and low technology. All other non-manufacturing products grouped as “not classified”.
- Sectoral datasets: constructed for export regressions and displacement effect regressions. Lists of reporter countries and their top 50 trading partners match the bilateral trade datasets. Sectoral trade values converted to volumes using the reporter country’s export (import) prices as deflators. Sectoral time trends included in regressions to address bias on sectors.
- Bilateral real exchange rate: defined as the bilateral nominal exchange rate divided by the relative CPI, with higher values denoting real depreciation of the reporter country’s currency.
- Other macro variables: export and import price deflators, exchange rates, CPI, and total domestic demand are from the IMF’s International Financial Statistics and World Economic Outlook (April 2010).
- Sample period: 1990 to 2009.
- Footnote: "We rank trading partners based on bilateral trade values (exports + imports) in 2009."

### Industry classification (Table A1)
- High-technology industries
  - Aircraft and spacecraft 353
  - Pharmaceuticals 2423
  - Office, accounting and computing machinery 30
  - Radio, TV and communications equipment 32
  - Medical, precision and optical instruments 33
- Medium-high-technology industries
  - Electrical machinery and apparatus, n.e.c. 31
  - Motor vehicles, trailers and semi-trailers 34
  - Chemicals excluding pharmaceuticals 24 excl. 2423
  - Railroad equipment and transport equipment, n.e.c. 352 + 359
  - Machinery and equipment, n.e.c. 29
- Medium-low-technology industries
  - Building and repairing of ships and boats 351
  - Rubber and plastics products 25
  - Coke, refined petroleum products and nuclear fuel 23
  - Other non-metallic mineral products 26
  - Basic metals and fabricated metal products 27-28
- Low-technology industries
  - Manufacturing, n.e.c.; Recycling 36-37
  - Wood, pulp, paper, paper products, printing and publishing 20-22
  - Food products, beverages and tobacco 15-16
  - Textiles, textile products, leather and footwear 17-19
- Source: the OECD’s Science, Technology and Industry Scoreboard (2005)

### Key summary statistics (Tables A2–A4 highlights)
- Table A3 (selected lines as presented)
  - Standard Deviation
    - Growth in Bilateral Export 0.0680 0.2869416
    - Growth in Bilateral Imports 0.0620 0.3839416
    - Growth in Domestic Demand in Reporter Countries 0.0190 0.039416
    - Growth in Domestic Demand in Partner Countries 0.0310 0.069326
    - Nominal Bilateral Exchange Rate -1.8172 2.5399967
    - Real Bilateral Exchange Rate -1.9482.3679967
    - Relative CPI 0.1310.9399967
  - Mean
  - Number of Obs
  - Notes: All variables are measured in logarithms. Bilateral exports and imports are deflated using reporter country’s aggregate export and import prices.
- Table A3. Unit Labor Cost, percent change between 2000 and 2009 (relative figures as presented)
  - Greece 37.6% relative to Germany 28.7% relative to Euro 12 14.9%
  - Ireland 34.0% relative to Germany 25.4% relative to Euro 12 11.9%
  - Italy 32.2% relative to Germany 23.7% relative to Euro 12 10.4%
  - Portugal 27.0% relative to Germany 18.8% relative to Euro 12 6.0%
  - Spain 31.4% relative to Germany 22.9% relative to Euro 12 9.7%
  - Germany 6.9% relative to Germany 0.0% relative to Euro 12 -10.8%
  - France 21.1% relative to Germany 13.3% relative to Euro 12 1.1%
  - Relative to Euro 12 excl Germany 8.4% 5.6% 4.1% 0.0% 3.5% -15.8% -4.6%
  - Source: Eurostat
- Table A4. Average Annual Real Labor Productivity Growth (2000–2007) (Relative to Euro Area Average)
  - Greece Market Economy 2.1%
  - Italy Market Economy -0.8%
  - Ireland Market Economy 2.8%
  - Portugal Market Economy 0.1%
  - Spain Market Economy 0.2%
  - Total Manufacturing excl electrical: Greece -2.9% Italy -3.9% Ireland 1.4% Portugal -2.6% Spain -2.6%
  - ICT (Electrical & telecommunication): Greece 1.7% Italy 0.0% Ireland 1.9% Portugal 2.5% Spain -1.2%
  - Construction: Greece 1.4% Italy -1.4% Ireland -1.8% Portugal -1.4% Spain -0.8%
  - Markets services: Greece 1.1% Italy -1.8% Ireland 2.5% Portugal -1.5% Spain -0.5%
    - Distribution: 1.7% -1.6% 0.1% -2.6% -0.9%
    - Finance: 3.7% 0.8% 4.7% 4.3% 5.8%
    - Personal services: -0.5% -3.5% 1.9% -2.9% -2.6%
  - Source: EU-KLEMS database

### Regression tables — selected coefficients and notes
- Table A5. Sectoral Export Regressions (dependent variable: annual bilateral sectoral export volume of 11 euro countries with their top 50 trading partners from 1990 to 2009)
  - Coefficients (examples preserved exactly as presented)
    - Nominal Exchange Rate 0.526*** 0.447*** 0.444*** 0.532*** 0.533*** 0.501*** 0.500***  
    - Relative CPI -0.605*** -0.519*** -0.516*** -0.613*** -0.615*** -0.585*** -0.584***
    - Real Exchange Rate 0.564***
    - Reporter Demand 0.392*** 0.368*** 0.370*** 0.383*** 0.391*** 0.388*** 0.403*** 0.406***
    - Partner Demand 1.302*** 1.318*** 1.243*** 1.253*** 1.182*** 1.181*** 1.401*** 1.402***
    - PartnerDemand X P 0.435*** 0.439*** 0.350*** 0.350*** -0.492*** -0.493***
    - X Low 0.0416*** 0.0415*** -0.0877*** -0.0877*** 0.0751*** 0.0753***
    - X Medium-Low 0.0290** 0.0290** -0.0404 -0.0404* 0.0494*** 0.0495***
    - X Medium-High 0.0407*** 0.0406*** 0.0411 0.0411 0.0989*** 0.0990***
    - X High 0.114*** 0.114*** 0.0523** 0.0523** 0.132*** 0.132***
  - Interaction examples (preserved)
    - PartnerDemand X R(R=GIPS) X P 0.704*** 0.364* 0.297**
    - PartnerDemand X R(R=Grc) X P -0.376* 0.952** -0.170
    - PartnerDemand X R(R=Prt) X P 2.635*** 1.013*** 0.224
    - PartnerDemand X R(R=Esp) X P 1.842*** -0.3160 0.644***
  - Observations and fit
    - Obs ervations4739647396473964739647396473964739647396
    - R-squared 0.8370.8370.8400.8460.8390.8390.8420.843
  - Significance notation: *** p<0.01, ** p<0.05, * p<0.1.
  - Notes: All regressions include a full set of country fixed effects and sectoral time trends. Independent variables include nominal exchange rate, relative CPI (reporter CPI/Partner CPI), real exchange rate, reporter demand, partner demand, and interactions of trading partner's domestic demand with various country pair dummies and sectoral dummies.
  - Example interpretation provided in notes: "For example, the coefficient of -0.442 for 'PartnerDemand X R(R=Grc) X P X Low' in column (4) indicates the marginal difference in demand elasticity for exports of low technology goods from Greece to CEE countries."
- Table A6. Sectoral Displacement Effect Regressions (dependent variable: annual bilateral sectoral import volume of 17 countries with their top 50 trading partners from 1990 to 2009)
  - Coefficients (examples preserved exactly as presented)
    - Nominal Exchange Rate 0.144* 0.153** 0.154** 0.149* 0.391* 0.438**
    - Relative CPI -0.285*** -0.290*** -0.291*** -0.291*** -0.211 -0.255
    - Real Exchange Rate 0.208***
    - Reporter Demand 0.933*** 0.963*** 0.880*** 0.876*** 0.914*** 0.450** 0.317*
    - Partner Demand 1.114*** 1.099*** 0.795*** 0.806*** 0.791*** 0.933*** 1.003***
    - Import from China 0.335*** 0.332** 0.332*** 0.0862* 0.0762**
    - Import from China X P(P=Euro) -0.179*** -0.179***; X Low 0.111*** 0.111***; X Medium-Low 0.0941*** 0.0941***; X Medium-High 0.134*** 0.134***; X High 0.194*** 0.194***
    - Import from China X P(P=GIPS) 0.134*** 0.155**; X Low -0.0988*** -0.100***; X Medium-Low -0.117*** -0.118***; X Medium-High -0.174*** -0.176***; X High -0.218*** -0.219***
    - Import from China X P(P=Grc) 0.199*** 0.0476 0.265***; X Low -0.193*** -0.0990*** -0.195**; X Medium-Low -0.191*** -0.111*** -0.191**; X Medium-High -0.389*** -0.275*** -0.391**; X High -0.374*** -0.209*** -0.375**
  - Observations and fit
    - Ob s erv ation s296052960529605296052960564956495
    - R-s qu ared0.7470.7460.7550.7590.7530.8730.900
  - Notes: The 17 countries: United States, United Kingdom, Austria, Belgium, Denmark, France, Germany, Italy, Netherlands, Sweden, Switzerland, Japan, Spain, Turkey, Russian Federation, Czech Republic, and Poland. "Import from China" is the reporter country's import from China. Column (7) and (8) use the sub-sample where trading partners are only euro countries. All regressions include a full set of country fixed effects and sectoral time trends. Significance notation: *** p<0.01, ** p<0.05, * p<0.1.
  - Example interpretation provided in notes: "For example, 0.111 for 'Import from China X P(P=Euro) X Low' in column (3) means the marginal effect of imports of low technology goods from China on the reporter country’s imports from euro countries."

*Source: _wp12236 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12236.pdf_