## wpiea2019093

## Source details

**Canonical URL:** [wpiea2019093](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019093.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019093.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019093.pdf.json)

---

### Context and scope
- New firm-level database covering firms in 12 Asian economies over 1994-2016 with comprehensive information on firms’ FX liabilities.
- Focus: effect of exchange rate movements and U.S. monetary policy on corporate vulnerability, conditional on foreign-currency leverage.
- Vulnerability metric: Altman’s Emerging Market Z’’-score (aggregate measure combining profitability, leverage, liquidity, and solvency). Lower Z’’-scores indicate higher probability of corporate bankruptcy.

### Key statistics and stylized facts
- Global and regional debt dynamics:
  - Global corporate debt more than doubled over the past decade.
  - In emerging Asia, corporate debt increased to about 114 percent of GDP in 2017:Q4 from 71 percent in 2007:Q4.
  - Total stock of US dollar-denominated debt of non-banks in emerging Asia increased to about USD 1.4 trillion in 2017:Q4 from USD 0.41 trillion in 2007:Q4.
- Sample and coverage:
  - 12 Asian economies: China, Hong Kong SAR, India, Indonesia, Korea, Malaysia, Philippines, Singapore, Sri Lanka, Taiwan POC, Thailand and Vietnam.
  - Year coverage: 1994 to 2016; Vietnam starts in 2005.
  - Resulting unbalanced panel: 200681 firm‑year observations, which includes about 18000 firms.
- Vulnerability and shock magnitudes:
  - A 30 percent domestic currency depreciation against the U.S. dollar is associated with a 0.4 decrease in the Z’’-score.
  - That 0.4 decline corresponds, on average, to a two‑notch downgrade in the corporate credit rating (e.g., from A to BBB+).
  - Such a shock will result in 7 percent of firms in the sample falling into D‑rating bucket (or bankruptcy).

### Altman EME Z''-score (definition and mapping)
- Z''-score formula:
  - Z'' = 3.25 + 6.56(X1) + 3.26(X2) + 6.72(X3) + 1.05(X4)
  - Definitions:
    - X1 = Working Capital / Total Assets (liquidity)
    - X2 = Retained Earnings / Total Assets (leverage)
    - X3 = Operating Income (EBIT) / Total Assets (profitability)
    - X4 = Total Equity / Total Assets (solvency)
- Z''-score to rating bands (ranges preserved):
  - > 8.5 : AAA
  - 7.60 - 8.15 : AA+
  - 7.30 - 7.60 : AA
  - 7.00 - 7.30 : AA-
  - 6.85 - 7.00 : A+
  - 6.65 - 6.85 : A
  - 6.40 - 6.65 : A-
  - 6.25 - 6.40 : BBB+
  - 5.85 - 6.25 : BBB
  - 5.65 - 5.85 : BBB-
  - 5.25 - 5.65 : BB+
  - 4.95 - 5.25 : BB
  - 4.75 - 4.95 : BB-
  - 4.50 - 4.75 : B+
  - 4.15 - 4.50 : B
  - 3.75 - 4.15 : B-
  - 3.20 - 3.75 : CCC+
  - 2.50 - 3.20 : CCC
  - 1.75 - 2.50 : CCC-
  - < 1.75 : D

### Data construction and measurement
- Currency decomposition:
  - Key contribution: firm‑level currency decomposition of Asian corporates’ debt including bond issuance and syndicated loans.
  - Issuance data used for USD debt because balance sheets do not consistently report proportion of USD debt.
- Data sources:
  - Thomson Reuters Datastream Worldscope (firm balance sheet information).
  - Thomson One Banker (decomposition of leverage by currency).
- Sample construction and cleaning:
  - Exclude observations with no total asset data and observations with noticeable error.
  - Winsorize by dropping outliers outside the 0.5-99.5 percentile for Z''-score, ROA or leverage.
- Notable micro facts:
  - Syndicated loans are a major source of external finance in EMEs and are often USD‑denominated.
  - Average maturity of foreign currency debt has decreased, while local currency debt maturity has increased.

### Empirical strategy (main regressions)
- Firm-level vulnerability regression (baseline):
  - Z_{i,j,t} = α + β1 F_{i,j,t} + β2 M_{j,t} + β3 S_{i,j,t} × ΔER_{j,t} + β4 S_{i,j,t} × ΔIR_t + μ_i + ε_{i,j,t}
  - Where ΔER_{j,t} = change in exchange rate (positive = appreciation), S = share of U.S. dollar debt, ΔIR = global/domestic interest rate measure, μ_i = firm fixed effect.
- Investment regression:
  - ΔY_{i,j,t} = α + β1 F_{i,j,t} + β2 M_{j,t} + β3 S_{i,j,t} × ΔER_{j,t} + μ_i + ε_{i,j,t}
  - ΔY denotes percentage changes in capital expenditures (CAPEX scaled by lagged assets).
- Fixed effects:
  - Firm fixed effects included; year and country‑year effects largely captured via controls (VIX and exchange rate changes). Robustness checks with other fixed effect combinations are similar.

### Baseline empirical findings
- Exchange rate effects on vulnerability:
  - Local currency depreciation against the U.S. dollar is significantly associated with increased corporate vulnerability (lower Z''-score).
  - Coefficient examples (firm FE regressions):
    - ∆ Nominal Exchange Rate: positive and significant (e.g., Column estimates 1.355*** (0.151); 1.171*** (0.156); 1.365*** (0.155) across specifications).
    - ∆ Real Exchange Rate: positive and significant (e.g., 1.267*** (0.144); 1.055*** (0.150); 1.285*** (0.149)).
    - ∆ NEER and ∆ REER: positive and significant but with smaller magnitudes (e.g., ∆ NEER 0.768*** (0.157); ∆ REER 0.523*** (0.153)).
- U.S. monetary policy:
  - ∆ U.S. Shadow Rate: strongly negatively associated with Z''-score (e.g., -0.0257*** (0.00320); -0.0260*** (0.00333)), indicating tightening U.S. monetary policy raises corporate vulnerability in Asia.
  - Changes in domestic short‑term interest rate: point estimates negative but statistically insignificant in main specifications.
- Global risk (VIX):
  - VIX enters negatively and significantly in many specifications (e.g., -0.00673*** (0.00145); -0.00724*** (0.00145)).

### Role of U.S. dollar debt (currency decomposition results)
- Subsample: firms that ever issued USD debt (N around 15,614 in many columns; subsample statistics reported).
- Interaction effects:
  - U.S. dollar Debt Portion × ∆ Exchange Rate: positive and statistically significant interaction coefficients indicate depreciation has a stronger adverse effect on Z''-scores for firms with higher USD debt shares (examples: interaction estimates 3.949*** (1.471); 3.322*** (0.985); 3.417** (1.524) in Table 8 columns).
  - U.S. dollar Debt Portion × ∆ Short-term Interest Rate: negative interaction coefficients in some columns (e.g., -0.114* (0.0581); -0.114** (0.0580)), indicating sensitivity to domestic rates for high USD debt firms in some specifications.
- Quantile/threshold evidence (Table 10):
  - USD debt share < 10 percent: competitiveness (trade) channel dominates.
  - USD debt share between 10 percent and 20 percent: financial channel offsets competitiveness channel.
  - USD debt share > 20 percent: financial channel dominates.
  - Quantile regression results by USD debt proportion bins show heterogeneous ∆ Nominal Exchange Rate coefficients across bins (e.g., Column (1) <1.5%: ∆ Nominal Exchange Rate -0.354** (0.123); Column (2) [1.5%,4.5%] -2.210*** (0.698); Column (6) >35.9% 3.613** (1.231)).

### Investment results: competitiveness versus financial channels
- Average effect:
  - Local currency depreciation is, on average, associated with investment expansion (competitiveness gains), but this average masks heterogeneity by USD debt exposure.
- Interaction with USD debt share:
  - As U.S. dollar Debt Portion increases, the positive impact of depreciation on investment declines and eventually becomes negative.
  - Interaction coefficients in investment regressions are positive and significant (e.g., U.S. dollar Debt Portion × ∆ Real Exchange Rate 0.0375** (0.0167); × ∆ Nominal Exchange Rate 0.0312** (0.0151)), implying marginal effect of depreciation on CAPEX depends on USD debt share.
- Quantile/threshold evidence on CAPEX (Table 13):
  - USD debt share <= 1.5%: competitiveness channel dominates (CAPEX responds positively to depreciation).
  - USD debt share between 1.5% and 20%: mixed/offsetting effects.
  - USD debt share > 20%: financial channel dominates and investment contracts with depreciation.
- Exporter vs non-exporter comparison (Table 14):
  - Non‑exporters with USD borrowing are more sensitive to exchange rate fluctuations (lack USD revenues for natural hedge).
  - U.S. dollar Debt Portion coefficients differ: non‑exporter -0.00250 (0.00186) vs exporter 0.0201** (0.0102) in one specification; interaction patterns differ across real/nominal specifications.

### Robustness and sensitivity
- Results robust across exchange rate measures: bilateral nominal, bilateral real, NEER, and REER (magnitudes smaller for NEER/REER and for real vs nominal bilateral measures).
- Results robust after controlling for VIX and other global factors and when using alternative fixed effect structures.

### Implications and policy recommendations
- Core policy implications:
  - Exchange rate management and corporate FX‑risk regulation matter: high shares of foreign‑currency debt make firms especially sensitive to currency depreciations.
  - Monitor cross‑border credit and currency composition of corporate debt for financial stability surveillance.
  - Use macroprudential and microprudential policies to limit build‑up of foreign exchange balance‑sheet exposures and contain excessive leverage increases.
  - Fill data gaps on corporate sector finances, including foreign currency exposures and natural hedging; data on derivatives hedging are not available.
  - Prepare for an increase in corporate failures with shifts in major advanced economies’ monetary policy and, where needed, reform corporate insolvency regimes.
  - Global policy coordination and domestic buffers are important given U.S. monetary policy spillovers.

### Summary conclusions
- Asia’s nonfinancial corporates remain vulnerable to tightening global financial conditions.
- Tightening of U.S. monetary policy and local currency depreciation vis‑à‑vis the U.S. dollar increase probability of corporate financial distress (lower Z’’-score).
- The effect of exchange rate depends on firms’ foreign currency liability:
  - USD debt share < 10 percent: competitiveness channel dominates.
  - USD debt share between 10 and 20 percent: channels offset.
  - USD debt share > 20 percent: financial channel dominates.
- Exchange rate depreciation increases firm‑level investment on average, but for firms with large FX liabilities investment contracts with depreciation; non‑exporters that borrowed in USD are more sensitive to exchange rate fluctuations.

*Source: wpiea2019093.*

### REFERENCES..............................................................................................................

### wpiea2019093 - REFERENCES

### Context and scope
- New firm-level database covering firms in 12 Asian economies over 1994-2016 with comprehensive information on firms’ FX liabilities.
- Focus: effect of exchange rate movements and U.S. monetary policy on corporate vulnerability, conditional on foreign-currency leverage.
- Vulnerability metric: Altman’s Emerging Market Z’’-score (aggregate measure combining profitability, leverage, liquidity, and solvency). Lower Z’’-scores indicate higher probability of corporate bankruptcy.

### Stylized facts and key statistics
- Global and regional debt dynamics:
  - Global corporate debt more than doubled over the past decade.
  - In emerging Asia, corporate debt increased to about 114 percent of GDP in 2017:Q4 from 71 percent in 2007:Q4.
  - Total stock of US dollar-denominated debt of non-banks in emerging Asia increased to about USD 1.4 trillion in 2017:Q4 from USD 0.41 trillion in 2007:Q4.
- Sample and periods referenced: comparisons include the period prior to the Asian financial crisis (AFC) and the post-crisis buildup through 2016–2017.
- Exchange rate effect magnitude:
  - A 30 percent domestic currency depreciation against the U.S. dollar is associated with a 0.4 decrease in the Z’’-score.
  - That 0.4 decline corresponds, on average, to a two-notch downgrade in the corporate credit rating (e.g., from A to BBB+).
  - Such a shock will result in 7 percent of firms in the sample falling into D-rating bucket (or falling into bankruptcy).

### Main empirical findings
- Aggregate corporate health
  - The Altman Z’’-score indicates Asia’s nonfinancial corporate sector is generally healthier than in the runup to the AFC.
  - Corporate vulnerabilities are higher in some economies, including Hong Kong SAR, Indonesia, the Philippines, and Singapore.
- Exchange rate as shock amplifier
  - Overall evidence shows exchange rate depreciation increases the probability of default of Asian firms.
  - Statistically and economically significant associations are found between local currency depreciation against the U.S. dollar (both nominal and real) and higher corporate sector vulnerability.
  - Results are robust when replacing bilateral exchange rates with NEER or REER and after controlling for other global factors such as the VIX.
- U.S. monetary policy
  - Changes in U.S. monetary policy rates have significant effects on Asia’s corporate vulnerability, with a tightening of U.S. monetary policy increasing corporate sector vulnerability in Asia.
- Role of foreign-currency debt (FX liability)
  - The impact of exchange rate depreciation on corporate vulnerability is conditional on the share of U.S. dollar-denominated debt on firms’ balance sheets.
  - Interaction and quantile regressions findings:
    - When the share of U.S. dollar debt is below 10 percent, the competitiveness (trade) channel dominates the financial channel.
    - When the share of U.S. dollar debt is between 10 and 20 percent, the financial channel offsets the competitiveness channel.
    - When the share of U.S. dollar debt is higher than 20 percent, the financial channel dominates.
  - Conclusion: high FX liability amplifies the effect of exchange rate movements—exchange rates act as shock amplifiers rather than shock absorbers.
- Investment dynamics
  - On average, exchange rate depreciation increases firm-level investment.
  - For firms with large FX liabilities, the relationship reverses and investment contracts with depreciation.
  - The debt-overhang channel (where FX liabilities offset competitiveness gains) is more pronounced for non-exporter firms.

### Methodology highlights
- Use of Altman Z’’-score to aggregate multiple firm-level indicators (profitability, leverage, liquidity, solvency) into a single vulnerability metric—preferred over univariate indicators for predicting corporate distress.
- Construction of a comprehensive dataset of currency decomposition of corporate debt that includes both bond issuance and syndicated loans (syndicated loans being a dominant source of foreign currency debt for firms of various sizes).
- Inclusion of pre-AFC period as benchmark to assess evolution of corporate vulnerabilities.

### Implications
- Exchange rate management and corporate FX-risk regulation matter: high shares of foreign-currency debt make firms especially sensitive to currency depreciations.
- Monitoring cross-border credit and currency composition of corporate debt is important for financial stability surveillance.
- Policy measures to mitigate FX-risk (e.g., hedging markets, currency matching of assets and liabilities, macroprudential limits on foreign-currency borrowing) could reduce the amplification channel.
- U.S. monetary policy spillovers to Asia underline the need for global policy coordination and domestic buffers to absorb external shocks.

*Source: wpiea2019093 - REFERENCES.*

### Section 3 describes the data construction and the regression models that we use to conduct

### wpiea2019093 - Section 3 describes the data construction and the regression models that we use to conduct

### II. THE ALTMAN EME Z’’-SCORE
- The Altman Z’’-score is an established indicator of corporate vulnerability in EMEs; developed by Edward Altman (1968) and updated for EMEs (Altman, 2005; Altman et al., 2016).
- Altman (2005) establishes a correspondence between the Z’’-score and corporate bond ratings in EMEs (Text Table 1).
- Z’’-score formula (as presented in source):
  - 푍푍′′ = 3.25+ 6.56(푋푋1) + 3.26(푋푋2) + 6.72(푋푋3) + 1.05(푋푋4)
  - Definitions (as provided):
    - 푋푋1 = measure of liquidity (working capital / total assets)
    - 푋푋2 = measure of leverage (pecking order theory basis)
    - 푋푋3 = measure of profitability (EBIT / total assets)
    - 푋푋4 = measure of solvency (how much assets can drop before insolvency)
- Rationale for the constant term:
  - Constant 3.25 derived from median Z’’-score for bankrupt U.S. firms (median of bankrupt U.S. subset without constant was -3.25; constant shifts formulation for standardization).
- Advantages of Z’’-score over single indicators:
  - Statistically dominates single ratios in predicting default probability; robust across model specifications (Altman et al., 2016).
  - Economically, single measures like leverage can reflect multiple underlying drivers (risky refinancing, investment demand, firm age), so Z’’-score aggregates multiple dimensions of vulnerability.
- Z’’-score to rating bands (excerpt from Text Table 1; ranges preserved as in source):
  - > 8.5 : AAA
  - 7.60 - 8.15 : AA+
  - 7.30 - 7.60 : AA
  - 7.00 - 7.30 : AA-
  - 6.85 - 7.00 : A+
  - 6.65 - 6.85 : A
  - 6.40 - 6.65 : A-
  - 6.25 - 6.40 : BBB+
  - 5.85 - 6.25 : BBB
  - 5.65 - 5.85 : BBB-
  - 5.25 - 5.65 : BB+
  - 4.95 - 5.25 : BB
  - 4.75 - 4.95 : BB-
  - 4.50 - 4.75 : B+
  - 4.15 - 4.50 : B
  - 3.75 - 4.15 : B-
  - 3.20 - 3.75 : CCC+
  - 2.50 - 3.20 : CCC
  - 1.75 - 2.50 : CCC-
  - < 1.75 : D

### III.A. Data — construction and coverage
- Key contribution: firm‑level currency decomposition of Asian corporates’ debt (including both bond issuance and syndicated loans).
- Data sources:
  - Thomson Reuters Datastream Worldscope (firm balance sheet information).
  - Thomson One Banker (decomposition of leverage by currency).
  - Worldscope and Thomson One Banker share firm identifiers to enable firm‑level merge.
- Sample coverage:
  - 12 Asian economies: China, Hong Kong SAR, India, Indonesia, Korea, Malaysia, Philippines, Singapore, Sri Lanka, Taiwan POC, Thailand and Vietnam.
  - Year coverage: 1994 to 2016; Vietnam starts in 2005.
  - Includes both active and inactive firms.
- Sample construction and cleaning:
  - Exclude observations with no total asset data and observations with noticeable error.
  - Winsorize by dropping outliers whose Z’’-score, return on assets (ROA) or leverage lie outside the 0.5-99.5 percentile of the whole sample.
  - Resulting unbalanced panel: 200681 firm‑year observations, which includes about 18000 firms.
- Rationale for using issuance data for USD debt:
  - Balance sheets do not consistently report proportion of USD debt; reporting is discretionary and subject to mismeasurement.
  - Issuance data (bonds and syndicated loans) better captures USD debt exposure.
- Additional empirical facts highlighted:
  - Syndicated loans are a major source of external finance in EMEs and are often USD‑denominated.
  - Micro data show average maturity of foreign currency debt has decreased, while local currency debt maturity has increased (figures referenced in source).

### III.B. Empirical methodology
- Main firm‑level vulnerability regression (equation 1):
  - 푍푍푤,푗,푐 = 훼 + 훽1 퐹퐹푤,푗,푐 + 훽2 푀푀푗,푐 + 훽3 푆푆푤,푗,푐 × 𝐸𝐸𝐸𝐸푗,푐 + 훽4 S푤,푗,푐 × 𝐼𝐼𝐸퐸푐 + 휇푤 + 휀푤,푗,푐
  - Components:
    - 푍푍푤,푗,푐: firm i Altman Z’’-score in country j at time t.
    - 퐹퐹: vector of firm‑level controls.
    - 푀푀: vector of macro/market controls for country j.
    - 𝐸𝐸𝐸𝐸푗,푐: change in exchange rate (positive change = appreciation).
    - 𝐼𝐼𝐸퐸푐: measure of global interest rates (or domestic short‑term interest rate).
    - 푆푆: share of U.S. dollar denominated debt on firm i’s balance sheet at year t.
    - 휇푤: firm fixed effect.
- Investment regression (equation 2):
  - Δ푌푌푤,푗,푐 = 훼 + 훽1 퐹퐹푤,푗,푐 + 훽2 푀푀푗,푐 + 훽3 푆푆푤,푗,푐 × 𝐸𝐸𝐸𝐸푗,푐 + 휇푤 + 휀푤,푗,푐
  - Δ푌푌 denotes percentage changes in capital expenditures (proxy for firm investment).
  - Regressions are also run separately for exporters and non‑exporters.
- Fixed effects:
  - Results presented with firm fixed effects; year fixed effects and country‑year fixed effects are captured by coefficients on VIX and country change in exchange rate in main specifications. Robustness checks with other fixed effect combinations are qualitatively similar.

### IV. EMPIRICAL RESULTS — Baseline findings
- Baseline: local currency depreciation against U.S. dollar is significantly associated with increased corporate vulnerability (lower Z’’-score).
- Magnitude example:
  - A 30 percent domestic currency depreciation against the U.S. dollar is associated with a 0.4 decrease in the Z’’-score.
  - This corresponds, on average, to a two‑notch downgrade in corporate credit rating (e.g., from A to BBB+).
  - Such a shock will result in 7 percent of firms in the sample falling into D‑rating bucket (or bankruptcy), per Text Table 1.
- Interest rate findings:
  - Changes in domestic short‑term interest rate: negative but insignificantly associated with corporate vulnerability.
  - Changes in U.S. shadow rate: strongly negatively associated with corporate vulnerability; association robust after controlling for VIX and bilateral exchange rate.
  - Interpretation: U.S. rates matter more for corporate vulnerability than domestic rates due to tighter USD supply, higher external financing costs, higher funding costs for global banks, and U.S. dollar appreciation increasing real debt burden for USD‑exposed firms.
- Robustness across exchange rate measures:
  - Results hold using bilateral nominal exchange rate, bilateral real exchange rate, NEER, and REER.
  - Magnitudes smaller for NEER/REER (less volatile) and for real vs nominal bilateral measures.

### IV.B. Corporate vulnerability — competitiveness versus financial channels
- Mechanisms:
  - Competitiveness (trade) channel: depreciation can boost exporters’ profits and reduce vulnerability.
  - Financial (balance sheet) channel: depreciation increases real burden of FX debt and raises vulnerability for FX‑exposed firms.
- Subsample focus: firms that ever had access to U.S. dollar denominated debt (to better identify balance‑sheet effects and avoid under‑coverage biases).
- Interaction model findings (Tables 8–9):
  - Local currency depreciation is significantly associated with corporate vulnerability after controlling for share of USD debt.
  - The marginal impact of depreciation is stronger for firms with higher share of USD‑denominated debt (interaction coefficient statistically significant).
- Quantile regression thresholds (Table 10):
  - When share of U.S. dollar debt is below 10 percent: competitiveness channel dominates (columns 1–3).
  - When share is between 10 percent and 20 percent: financial channel offsets competitiveness channel (column 4).
  - When share is higher than 20 percent: financial channel dominates (column 6).
- Conclusion: high FX liability amplifies the impact of exchange rate on corporate vulnerability — exchange rate fluctuations act as shock amplifiers, not absorbers.

### IV.C. Corporate investment — competitiveness versus financial channels
- Investment regressions (Tables 11–13 and comparisons):
  - Local currency depreciation is, on average, associated with investment expansion (reflecting competitiveness gains for exporters).
  - However, as the share of U.S. dollar‑denominated debt increases, the positive impact of depreciation on investment declines and eventually becomes negative (balance sheet channel dominates).
- Quantile regression thresholds for investment (Table 13):
  - USD debt share < 10 percent: competitiveness channel dominates (columns 1–3).
  - USD debt share between 10 and 20 percent: financial channel offsets competitiveness channel (column 4).
  - USD debt share > 20 percent: financial channel dominates (column 6).
- Exporter vs non‑exporter comparison (Table 14):
  - Local currency appreciation impact is negative for both exporters and non‑exporters, more significantly negative for exporters.
  - Marginal effect of exchange rate appreciation conditional on USD debt is more negatively significant for non‑exporters — non‑exporters with USD borrowing are more sensitive to exchange rate fluctuations because they lack natural hedges (no USD revenue).
- Robustness:
  - Results hold when using nominal exchange rate instead of real exchange rate.
- Overall conclusion from investment analysis: high FX liability amplifies exchange rate effects on investment; exchange rate fluctuations act as shock amplifiers.

### V. Conclusion and policy implications
- Database and scope:
  - Constructed a new firm‑level database for 12 Asian economies over 1994–2016 to assess corporate vulnerability and channels of global financial condition transmission.
- Key empirical conclusions:
  - Asia’s nonfinancial corporates remain vulnerable to tightening global financial conditions.
  - Tightening of U.S. monetary policy and local currency depreciation vis‑à‑vis the U.S. dollar increase probability of corporate financial distress (lower Z’’-score).
  - The effect of exchange rate depends on firms’ foreign currency liability:
    - USD debt share < 10 percent: competitiveness channel dominates.
    - USD debt share between 10 and 20 percent: channels offset.
    - USD debt share > 20 percent: financial channel dominates.
  - Exchange rate depreciation increases firm‑level investment on average, but for firms with large FX liabilities investment contracts with depreciation.
  - Non‑exporters that borrowed in USD are more sensitive to exchange rate fluctuations.
  - Overall: high FX liability amplifies exchange rate shocks for corporate sector.
- Policy recommendations (as presented):
  - Monitor vulnerable firms, especially systemically important ones, and banks/other sectors closely linked to them.
  - Fill data gaps on corporate sector finances, including foreign currency exposures and natural hedging.
  - Use macroprudential and microprudential policies to limit build‑up of foreign exchange balance sheet exposures and contain excessive leverage increases.
  - Prepare for an increase in corporate failures with shifts in major advanced economies’ monetary policy and, where needed, reform corporate insolvency regimes.
- Note: mitigation via derivatives is possible but data on derivatives hedging are not available.

*Source: wpiea2019093 - Section 3 describes the data construction and the regression models that we use to conduct*

### REFERENCES

### wpiea2019093 - REFERENCES

### References
- Acharya Viral, Stephen G. Cecchetti, José De Gregorio, Şebnem Kalemli-Özcan, Philip R. Lane, and Ugo Panizza, 2015. “Corporate Debt in Emerging Economies: A Threat to Financial Stability?”, Committee for International Policy Reform, Brookings Institution, Washington DC.
- Acharya, Viral, Eisert, Tim, Eufinger Christian and Hirsch, Christian (2015), “Real Effects of the Sovereign Debt Crisis in Europe: Evidence from Syndicated Loans”, Working Paper.
- Aguiar, M., 2005, “Investment, devaluation, and foreign currency exposure: The case of Mexico,” Journal of Development Economics, Vol. 78(1), pp.95-113.
- Alfaro, Laura, Gonzalo Asis, Anusha Chari, and Ugo Panizza (2017), “Lessons Unlearned? Corporate Debt in Emerging Markets” NBER Working Paper No. 23407.
- Alter, Adrian, and Elekdag, Selim (2016) “Emerging Market Corporate Leverage and Global Financial Conditions”, IMF Working Paper.
- Altman, Edward (1968), “Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy”, Journal of Finance, Sep. 1968.
- Altman, Edward (2005), “An Emerging Market Credit Scoring System for Corporate Bonds,” Emerging Market Review 6: 3011-323.
- Altman, Edward, Iwanicz‐Drozdowska, Małgorzata, Laitinen, Erkki and Suvas, Arto (2016), “Financial Distress Prediction in an International Context: A Review and Empirical Analysis of Altman's Z‐Score Model”. Journal of International Financial Accounting and Management, Volume 28, Issue 2, June 2017, Pages 131-171.
- Amiti, M, O Itskhoki, and J Konings (2012), "Importers, Exporter, and Exchange Rate Disconnect," CEPR Discussion Paper No. 9260.
- Avdjiev, Stefan, Bruno, Valentina, Koch, Catherine, and Hyun Song Shin (2018), “The Dollar Exchange Rate as a Global Risk Factor: Evidence from Investment”, BIS Working Papers, No 695.
- Bleakly, Hoyt and Cowan, Kevin (2008) “Corporate Dollar Debt and Depreciation: Much ado about Nothing?” The Review of Economics and Statistics, MIT Press, vol. 90(4), pages 612-626, November.
- Brauning, Falk and Victoria Ivashina (2017). Monetary Policy and Global Banking. NBER Working Paper No. 23316. Issued in April 2017.
- Bruno, Valentina and Hyun Song Shin (2015), “Capital Flows and the Risk-Taking Channel of Monetary Policy”, Journal of Monetary Economics, Volume 71.
- Bruno, Valentina and Hyun Song Shin (2017), “Global Dollar Credit and Carry Trades: A Firm-level Analysis”, Review of Financial Studies, vol 30, no 3, 2017, pp 703-49.
- Bruno, Valentina and Hyun Song Shin (2018), “Currency depreciation and emerging market corporate distress”, BIS Working Papers | No 753.
- Dao, Mai, Camelia Minoiu, and Jonathan D.Ostry, 2017, Corporate Investment and the Real Exchange Rate, IMF Working Paper WP/17/183.
- Hofmann, Boris, Shim, Ilhyock and Shin, Hyun Song (2017), “Sovereign Yields and the Risk-Taking Channel of Currency Appreciation”, BIS Working Papers, No 538.
- Ivashina, Victoria and Becker, Bo (2014), “Cyclicality of Credit Supply: Firm Level Evidence”, Journal of Monetary Economics, Volume 62, March 2014, Pages 76-93.
- McCauley, Robert, Patrick McGuire and Vladyslav Sushko, 2015, Global dollar credit: links to US monetary policy and leverage, BIS Working Papers No 483.
- Myers, Stwart and Majluf, Nicholas, 1984. “Corporate Financing and Investment Decisions when Firms Have Information that Investors Do Not Have.”, Journal of Financial Economics, Volume 13, Issue 2, June 1984, Pages 187-221.
- Powell, Jerome, 2017, Prospects for Emerging Market Economies in a Normalizing Global Economy, Speech, At the 2017 Annual Membership Meeting of the Institute of International Finance, Washington, D.C.
- Serena, Jose Maria and Sousa, Ricardo, 2017. “Does exchange rate depreciation have contractionary effects on firm-level investment?”, BIS Working Paper.
- Shin, Hyun Song, 2013. "The second phase of global liquidity and its impact on emerging economies," Proceedings, Federal Reserve Bank of San Francisco, issue Nov, pages 1- 10.
- Shin, Hyun Song and L Zhao (2013) “Firms as surrogate intermediaries: evidence from emerging economies”, Working Paper, December 2013.
- Wu, Jing, and Xia, Fan (2016), “Measuring the Macroeconomic Impact of Monetary Policy at the Zero Lower Bound”, Journal of Money, Credit and Banking, Vol. 48, No. 2–3.

### Appendix I. Figures — captions, sources, and key numeric labels
- Figure 1. Nonfinancial Corporate Credit, 1996Q4-2017Q4
  - Y-axis labels shown: 0, 50, 100, 150, 200, 250
  - Series/countries listed: Indonesia, Korea, Malaysia, Thailand, Australia, China, Hong Kong SAR, India, Japan, New Zealand, Singapore, EM Asia, EMs, All countries, AEs
  - Year markers: 1996:Q4, 2007:Q4, 2017:Q4
  - Unit: Nonfinancial Corporations Credit (Percent of GDP)
  - Source: Bank for International Settlements.
  - Note: 1/ EMs and All Countries data as of 2008:Q1. AEs: AdvancedEconomies; EMs: Emerging Markets; EM Asia: Emerging Markets in Asia.

- Figure 2: U.S. Monetary Policy Conditions and Cross-border Bank Credit
  - Panel (A): y-axis is the change in cross-border lending as a percentage of GDP lagged by one period; x-axis is the percentage change of actual Fedral Fund Rate.
  - Panel (B): y-axis is the total level of cross-border lending on the country’s balance sheet divided by the GDP lagged by one period; x-axis is the level of Federal Fund Rate.
  - Source: Authors’ caculation based on IMF’s International Investment Positions. Cross-border lending is the “Debt Instruments” from foreign deposit-taking institutes.

- Figure 3. Comparing Corporate Vulnerability of Asia, Z’’-score: 1996-2016

- Figure 4. Comparing Corporate Vulnerability of Asia, ICR: 1996-2016
  - ICR by Ranges in Asia: 1996, 2016 — range bins shown: <2, [2,5], [5,10), >10
  - Z-Score by Zones in Asia: 1996 vs 2016 — zones: Safe Zone, Grey Zone, Distress Zone

- Figure 5. Dynamics of Z’’-score, 1994-2016
  - The figures plot evolution of the Z’’-score and its four components in 12 Asian economies over 1995-2015.
  - Upper cap and lower cap represent the 90th and 10th percentile in a given year; red dots represent the mean in a given year.

- Figures 6–9. Dynamics of Z’’-score Components (1994–2016)
  - Figure 6. Dynamics of X1: Working Capital/Total Asset (Liquidity)
  - Figure 7. Dynamics of X2: Retained Earning/Total Assets (Leverage)
  - Figure 8. Dynamics of X3: Operating Income/Total Assets (Profitability)
  - Figure 9. Dynamics of X4: Total Equity/Total Liabilities (Solvency)

- Figure 10. Bond Issuances in Asia—1994-2017
  - Source: Thomson One Banker.
  - Note: Each bar represents the amount of total issuance in the 12 Asian economies in a given year.
  - Legend/series: Non-USD Bond, USD Bond
  - Y-axis unit: U.S. dollar Billion
  - Year range shown: 1994–2017

- Figure 11. Corporate Loans in Asia—1994-2017
  - Source: Thomson One Banker.
  - Note: Each bar represents the amount of total issuance in the 12 Asian economies in a given year.
  - Legend/series: Non-USD Loan, USD Loan
  - Y-axis unit: U.S. dollar Billion
  - Year range shown: 1994–2017

- Figure 12: Maturity Dynamics of Syndicated Loan Issuance: Asia 1994-2017
  - Source: Thomson One Banker.
  - Y-axis is the weighted (by dollar amount) average of syndicated loan maturity issued in a given year.

- Figure 13: Maturity Dynamics of Bond Issuance: Asia 1994-2017
  - Source: Thomson One Banker.
  - Y-axis is the weighted (by dollar amount) average of corporate bond maturity issued in a given year.
  - Two series shown: Non-USD Bond, USD Bond
  - Years (weighted by amount) plotted include 1993 through 2017.
  - Numeric ticks shown in examples: 0.00, 2.00, 4.00, 6.00, 8.00, 10.00, 12.00, 14.00; and 0, 2, 4, 6, 8, 10, 12, 14, 16, 18 across figures.

- Figure 14: Distribution of Corporates’ U.S. dollar Debt/Total Assets: Exporter/Non-exporter and Tradable sector/ Non-tradable sector
  - Source: Datastream Worldscope.
  - Classification rules:
    - A company is classified as an exporter if its Exports (WC07161) is nonnegative.
    - A company is classified as in tradable sector if its SIC-code (WC07202) is between 2000 and 3999.
  - Note: The categorizations are based on Du and Schreger (2016) and the results are similar to their findings.

*Source: wpiea2019093 - REFERENCES*

### APPENDIX II. TABLES

### APPENDIX II. TABLES

### Definitions and data sources (Table 1)
- Delta Nominal U.S. Exchange Rate: (Nominal U.S. exchange rate(t)-Nominal U.S. exchange rate(t-1))/Nominal U.S. exchange rate(t-1). Source: IMF WEO.
- Delta Real U.S. Exchange Rate: (Real U.S. exchange rate(t)-Real U.S. exchange rate(t-1))/Real U.S. exchange rate(t-1). Source: IMF WEO.
- Delta NEER: (NEER(t)-NEER(t-1))/NEER(t-1). Source: IMF WEO.
- Delta REER: (REER(t)-REER(t-1))/REER(t-1). Source: IMF WEO.
- Delta U.S. Shadow Rate: (U.S. Shadow Rate(t)-U.S. Shadow Rate(t-1))/U.S. Shadow Rate(t-1). Source: Wu and Xia (2016).
- ROA: Net Income/Total Assets. Source: Worldscope.
- Leverage: Total Liability/Total Asset. Source: Worldscope.
- Tobin's Q: Total Market Value of firm/Total Book Value of firm. Source: Worldscope.
- Cash: Cash Holdings. Source: Worldscope.
- U.S. dollar Debt Portion: (U.S. dollar loan+U.S. dollar bond)*Exchange Rate/Total Asset. Source: Calculation based on Thomson One and Worldscope.
- Z''-score: Altman's 2005 Z'' score for emerging market firms: Z''=3.25+6.56(X_1 )+3.26(X_2 )+6.72(X_3 )+1.05(X_4). Source: Calculation based on Worldscope database.
  - X1: Working Capital/Total Asset. Source: Worldscope.
  - X2: Retained Earnings/Total Assets. Source: Worldscope.
  - X3: Operating Income (EBIT)/Total Assets. Source: Worldscope.
  - X4: Total Equity/Total Assets. Source: Worldscope.
- Exporter: Dummy variable which equals 1 if the firm has exporting income on balance sheet. Source: Worldscope.

### Summary statistics (Table 2 A: overall sample; Table 2 B: subsample)
- Table 2 A (Overall sample; N and moments):
  - Z''-score: N 200681; Mean 7.113; St. Dev 5.224; Median 6.564.
  - ROA: N 200681; Mean .044; St. Dev .141; Median .049.
  - EBIT: N 200681; Mean .05; St. Dev .162; Median .059.
  - Leverage: N 200681; Mean .468; St. Dev .265; Median .455.
  - CAPEX: N 199116; Mean .054; St. Dev .195; Median .032.
  - Ln[Total Assets]: N 200681; Mean 5.694; St. Dev 2.245; Median 5.365.
  - Retained Earnings: N 200681; Mean .049; St. Dev .574; Median .126.
  - Working Capital: N 200681; Mean .178; St. Dev .278; Median .171.
  - Cash: N 163817; Mean .11; St. Dev .125; Median .07.
  - Employees: N 136938; Mean 5862.416; St. Dev 24453.73; Median 1085.
  - Tangibility: N 200256; Mean .319; St. Dev .222; Median .289.
  - Tobin’s Q: N 185191; Mean 1.245; St. Dev 11.524; Median .644.
  - Total Equity: N 200681; Mean .507; St. Dev .269; Median .516.

- Table 2 B (Subsample; N and moments):
  - Z''-score: N 15614; Mean 5.881; St. Dev 8.204; Median 5.775.
  - ROA: N 21493; Mean .054; St. Dev .123; Median .054.
  - EBIT: N 21349; Mean .061; St. Dev .155; Median .065.
  - Leverage: N 21463; Mean .426; St. Dev 1.327; Median .402.
  - CAPEX: N 19960; Mean .077; St. Dev .181; Median .042.
  - Ln[Total Assets]: N 21493; Mean 7.732; St. Dev 2.549; Median 7.462.
  - Retained Earnings: N 16994; Mean .042; St. Dev .755; Median .119.
  - Working Capital: N 19923; Mean .09; St. Dev .345; Median .099.
  - Cash: N 20365; Mean .072; St. Dev .089; Median .042.
  - Employees: N 15374; Mean 10670.77; St. Dev 34910.91; Median 2828.5.
  - Tangibility: N 21461; Mean .388; St. Dev .239; Median .38.
  - Tobin’s Q: N 20584; Mean 1.025; St. Dev 1.436; Median .756.
  - Total Equity: N 21493; Mean .409; St. Dev .348; Median .421.
  - USD bank loan: N 8052; Mean 643.476; St. Dev 1128.363; Median 150 (face value in million US).
  - USD bond: N 2872; Mean 522.151; St. Dev 1616.578; Median 300 (face value in million US).
  - Total USD debt: N 9448; Mean 0.316; St. Dev 1.879; Median 0.286.

- Additional notes:
  - Leverage is total liabilities divided by lagged total assets.
  - CAPEX, EBIT, Retained Earnings, Working Capital, Cash, Total Equity and Total USD debt are all scaled by lagged total assets.
  - Employees is the absolute number of employees of the firms.
  - USD bank loan and USD bond are the face value of the firms’ bond and bank loan issuance in million US.
  - Tangibility is the book value of Properties, Plants and Equipment (PPE) divided by lagged total assets.
  - Ln[Total Assets] is the natural log of total assets converted to USD.

### Baseline results: Exchange rate and firm Z''-score (Tables 4–7)
- General specification: regress firm-level Z''-score on percentage changes in exchange rates (nominal, real, NEER, REER), domestic short-term interest rate, U.S. shadow rate, and VIX. Country-time fixed effects absorbed into changes in nominal exchange rate and short-term interest rate. Firm fixed effects included.

- Table 4: Nominal Exchange Rate (columns (1)-(5)):
  - ∆ Nominal Exchange Rate:
    - Column (1) coefficient 1.355*** (0.151).
    - Column (3) coefficient 1.171*** (0.156).
    - Column (5) coefficient 1.365*** (0.155).
  - ∆ Short-term Interest Rate: column (2) coefficient -0.000332 (0.00249); column (4) coefficient -0.00190 (0.00254).
  - ∆ U.S Shadow Rate: column (3) coefficient -0.0257*** (0.00320); column (5) coefficient -0.0260*** (0.00333).
  - VIX: column (4) coefficient -0.00673*** (0.00145); column (5) coefficient -0.00431*** (0.00147).
  - Constant values: 7.117*** (0.00808), 7.129*** (0.00856), 7.069*** (0.00863), 7.204*** (0.0299), 7.216*** (0.0305).
  - Observations: 200,681; 187,726; 184,888; 184,888; 187,726 (by column).
  - R-squared: 0.571; 0.561; 0.578; 0.578; 0.561.
  - Firm FE: Yes.
  - Standard errors in parentheses. Significance: *** p<0.01, ** p<0.05, * p<0.1.

- Table 5: Real Exchange Rate (columns (1)-(5)):
  - ∆ Real Exchange Rate:
    - Column (1) coefficient 1.267*** (0.144).
    - Column (3) coefficient 1.055*** (0.150).
    - Column (5) coefficient 1.285*** (0.149).
  - ∆ Short-term Interest Rate: same reported values as Table 4 (-0.000332; -0.00160 with SEs).
  - ∆ U.S. Shadow Rate: -0.0257*** (0.00320); -0.0258*** (0.00334).
  - VIX: -0.00724*** (0.00145); -0.00465*** (0.00146).
  - Constant values and Observations/R-squared/Firm FE same as Table 4.

- Table 6: NEER (columns (1)-(5)):
  - ∆ NEER:
    - Column (1) coefficient 0.768*** (0.157).
    - Column (3) coefficient 0.694*** (0.168).
    - Column (5) coefficient 0.751*** (0.168).
  - ∆ Short-term Interest Rate: same small negative point estimates reported.
  - ∆ U.S. Shadow Rate: -0.0257*** (0.00320); -0.0284*** (0.00331).
  - VIX: -0.00641*** (0.00149); -0.00362** (0.00149).
  - Constant, Observations, R-squared, Firm FE same as above.

- Table 7: REER (columns (1)-(5)):
  - ∆ REER (labelled NEER in table header but context indicates REER results):
    - Column (1) coefficient 0.523*** (0.153).
    - Column (3) coefficient 0.394** (0.163).
    - Column (5) coefficient 0.464*** (0.162).
  - ∆ U.S. Shadow Rate: -0.0257*** (0.00320); -0.0290*** (0.00331).
  - VIX: -0.00713*** (0.00148); -0.00413*** (0.00149).
  - Constant, Observations, R-squared, Firm FE same as above.

### Currency decomposition: role of U.S. dollar debt (Tables 8–9)
- Sample: subsample of firms that have ever issued USD debt. Dependent variable: firm-level Z''-score. Main independent: U.S. dollar Debt Portion (scaled by lagged total assets). Firm fixed effects controlled.

- Table 8: Currency Decomposition—Nominal Exchange Rate (columns (1)-(10)):
  - ∆ Nominal Exchange Rate coefficients (various columns):
    - Column (1) 2.369*** (0.805).
    - Column (3) 1.329 (0.893).
    - Column (4) 1.219** (0.605).
    - Column (5) 1.273 (0.915).
    - Column (6) 2.339*** (0.819).
    - Column (7) 2.151*** (0.822).
    - Column (8) 2.469*** (0.829).
    - Column (9) 2.109*** (0.545).
  - U.S. dollar Debt Portion main effect (selected columns):
    - Column (1) -0.0450 (0.0886).
    - Column (2) -0.0161 (0.0394).
    - Column (4) -0.0455 (0.0906).
    - Column (5) -0.165* (0.0992).
    - Column (6) -0.0829* (0.0444).
  - Interaction U.S. dollar Debt Portion * ∆ Nominal Exchange Rate:
    - Column (4)-(6) interactions reported:
      - Column (4) 3.949*** (1.471).
      - Column (5) 3.322*** (0.985).
      - Column (6) 3.417** (1.524).
  - ∆ U.S. Shadow Rate coefficients and interactions:
    - ∆ U.S. Shadow Rate: -0.111*** (0.0165) in a column; -0.104*** (0.0167) in another; -0.0863*** (0.0169); -0.0989*** (0.0174) in later columns.
    - Interaction U.S. dollar Debt Portion * ∆ U.S. Shadow Rate: coefficients -0.0296 (0.0564) and -0.0316 (0.0564) (not significant).
  - VIX: negative and significant in several columns; e.g., -0.0213*** (0.00683); -0.0265** (0.0103); -0.0267*** (0.0103); -0.0212*** (0.00683).
  - Interaction U.S. dollar Debt Portion * ∆ Short-term Interest Rate: -0.114* (0.0581); -0.114** (0.0580) in columns (7)-(8).
  - Observations by column: 15,614; 15,047; 14,545; 15,614; 14,545; 15,047; 15,047; 15,046; 14,545; 14,545.
  - R-squared across columns: approx 0.304–0.307.
  - Firm FE: Yes.

- Table 9: Currency Decomposition—Real Exchange Rate (columns (1)-(11)):
  - ∆ Real Exchange Rate coefficients (selected):
    - Column (1) 1.911** (0.783).
    - Column (5) 1.204 (0.877).
    - Column (6) 1.153 (0.856).
    - Column (7) 1.076 (0.895).
    - Column (8) 1.921** (0.796).
    - Column (9) 1.764** (0.798).
    - Column (10) 1.925** (0.807).
    - Column (11) 1.693** (0.812).
  - U.S. dollar Debt Portion main effect (selected):
    - Column (1) -0.0468 (0.0886); Column (2) -0.0210 (0.0904); Column (3) -0.0455 (0.0906); Column (4) -0.0180 (0.0920).
    - Columns (5)-(7): -0.0946 (0.0925); -0.0806 (0.0948); -0.0671 (0.0944).
    - Columns (8)-(11): 0.00737 (0.0914); 0.00631 (0.0913); -0.0378 (0.0968); -0.0377 (0.0968).
  - Interaction U.S. dollar Debt Portion * ∆ Real Exchange Rate:
    - Columns (5)-(7) interactions: 3.003* (1.680); 2.334 (1.711); 2.980* (1.710).
  - Interaction U.S. dollar Debt Portion * ∆ Short-term Interest Rate:
    - Columns show -0.114** (0.0581) and -0.115** (0.0580).
  - VIX: negative and significant in some columns; e.g., -0.0260** (0.0105); -0.0274*** (0.0103); -0.0274*** (0.0103); -0.0268** (0.0105).
  - Observations by column: 15,614; 15,047; 14,546; 14,546; 15,614; 14,546; 15,047; 15,047; 15,047; 14,546; 14,546.
  - R-squared across columns: approx 0.304–0.306.
  - Firm FE: Yes.

### Corporate vulnerability: quantile regressions (Table 10)
- Dependent: Z''-score by USD Debt Proportion quantiles.
- Columns correspond to USD Debt Proportion bins and sample sizes; reported R-squared for bins:
  - Column (1) USD Debt Proportion <1.5%: Observations 10,579; R-squared 0.433.
  - Column (2) [1.5%, 4.5%]: Observations 1,381; R-squared 0.831.
  - Column (3) (4.5%, 10.8%]: Observations 2,081; R-squared 0.667.
  - Column (4) (10.8%, 20.3%]: Observations 1,789; R-squared 0.571.
  - Column (5) (23.6, 35.9]: Observations 837; R-squared 0.792.
  - Column (6) >35.9%: Observations 1,482; R-squared 0.745.
- Selected coefficient highlights (standard errors in parentheses):
  - USD debt portion: Column (2) -10.54* (6.046); Column (4) 12.02* (4.981); Column (5) 4.722* (2.639); Column (6) 6.0516 (10.136).
  - ∆ Nominal Exchange Rate: Column (1) -0.354** (0.123); Column (2) -2.210*** (0.698); Column (3) -2.207** (0.763); Column (5) 3.236* (1.487); Column (6) 3.613** (1.231).
  - VIX: Column (1) -0.0288* (0.0154); Column (3) -0.0618*** (0.0181); Column (4) -0.0417** (0.0189); Column (5) -0.0540** (0.0274); Column (6) -0.0966** (0.0385).
  - Constant values vary by column (e.g., 6.638*** (0.322) in col (1)).

### Investment regressions: exchange rates, USD debt, and firm investment (Tables 11–14)
- General: Investment = capital expenditure divided by lagged total assets. Subsample: firms with USD debt proportion measures (1420 firms).

- Table 11: Investment—Real Exchange Rate (columns (1)-(5)):
  - U.S. dollar Debt Portion: coefficients range from -0.000385 (0.00140) to -0.00197 (0.00160) (not significant).
  - ∆ Real Exchange Rate:
    - Column (1) -0.0379** (0.0166).
    - Column (2) -0.0462*** (0.0170).
    - Column (3) -0.0492*** (0.0170).
    - Column (4) -0.0502*** (0.0173).
    - Column (5) -0.0579*** (0.0178).
  - Interaction U.S. dollar Debt Portion * ∆ Real Exchange Rate:
    - Reported positive and significant: 0.0375** (0.0167); 0.0378** (0.0167); 0.0353** (0.0169); 0.0377** (0.0170).
  - Controls:
    - Leverage: -0.00532*** (0.000988); -0.00523*** (0.00103); -0.00505*** (0.00105).
    - Tobin’s Q: 0.0120*** (0.00106); 0.0116*** (0.00108).
    - Cash: 0.0682*** (0.0215) in one specification.
    - ∆ Short-term Interest Rate: small positive ~0.000459 with SE ~0.000340.
  - Observations: 19,202; 19,202; 19,202; 18,395; 17,527.
  - R-squared: 0.166; 0.166; 0.168; 0.163; 0.163.
  - Firm FE: Yes.

- Table 12: Investment—Nominal Exchange Rate (columns (1)-(5)):
  - U.S. dollar Debt Portion: coefficients from -0.000309 (0.00138) to -0.00110 (0.00144) (not significant).
  - ∆ Nominal Exchange Rate:
    - Column (1) -0.0257 (0.0167).
    - Column (2) -0.0281* (0.0169).
    - Column (3) -0.0357** (0.0172).
    - Column (4) -0.0458*** (0.0175).
    - Column (5) -0.0520*** (0.0181).
  - Interaction U.S. dollar Debt Portion * ∆ Nominal Exchange Rate:
    - 0.0312** (0.0151); 0.0297** (0.0152); 0.0307** (0.0153).
  - Controls:
    - Leverage: -0.00526*** (0.00103); -0.00509*** (0.00105).
    - Tobin’s Q: 0.0120*** (0.00106); 0.0117*** (0.00108).
    - Cash: 0.0672*** (0.0215).
    - ∆ Short-term Interest Rate: small positive ~0.000470 (0.000340).
  - Observations: 19,960; 19,202; 19,202; 18,395; 17,527.
  - R-squared: 0.157; 0.166; 0.166; 0.163; 0.163.
  - Firm FE: Yes.

- Table 13: Investment—Quantile Regressions (CAPEX by USD debt proportion bins)
  - Panels by USD Debt Proportion bins with Observations and R-squared:
    - <=1.5%: Observations 10,579; R-squared 0.157.
    - (1.5%, 4.5%]: Observations 1,381; R-squared 0.707.
    - (4.5%, 10.8%]: Observations 2,081; R-squared 0.653.
    - (10.8%, 20.3%]: Observations 1,789; R-squared 0.639.
    - (20.3, 35.9]: Observations 837; R-squared 0.681.
    - >35.9%: Observations 1,482; R-squared 0.526.
  - Selected coefficients (standard errors in parentheses):
    - USD debt portion: ranges from 1.207 (1.253) in col (1) to -0.640 (0.561) in col (6); only col (3) 0.257* (0.134) is marginally significant.
    - ∆ Nominal Exchange Rate: col (1) -0.0638** (0.0318); col (2) -0.0590** (0.0243); col (3) -0.0724** (0.0298); col (4) 0.0408 (0.0396); col (5) 0.0107* (0.00504); col (6) 0.0243** (0.00841).
    - Leverage: significant negative across bins (e.g., -0.00440** (0.00171) in col (1); -0.0118*** (0.00220) in col (3)).
    - Tobin’s Q: positive and significant across all bins (e.g., 0.00870*** (0.00141) in col (1); 0.0392*** (0.00473) in col (3)).
    - Cash: positive in several bins (e.g., 0.0744** (0.0332) in col (1); 0.153*** (0.0483) in col (4)).

- Table 14: Investment — Exporters vs. Non-exporters (columns (1)-(4))
  - Sample split: Exporter No (columns (1),(3)); Exporter Yes (columns (2),(4)).
  - Selected coefficients:
    - U.S. dollar Debt Portion:
      - Column (1) (Non-exporter) -0.00250 (0.00186).
      - Column (2) (Exporter) 0.0201** (0.0102).
      - Column (3) -0.00202 (0.00178).
      - Column (4) 0.0215** (0.0104).
    - ∆ Real Exchange Rate:
      - Column (1) -0.0427 (0.0260) (non-significant).
      - Column (2) -0.0303** (0.0124) (significant for exporters).
    - Interaction U.S. dollar Debt Portion * ∆ Real Exchange Rate:
      - Column (2) 0.0375** (0.0181) (Exporter sample).
      - Column (4) 0.0319 (0.0442) (less precise).
    - Interaction U.S. dollar Debt Portion * ∆ Nominal Exchange Rate:
      - Column (3) -0.0390* (0.0226) (non-exporters).
      - Column (4) 0.0331** (0.0166) (exporters).
    - Leverage: negative and significant across both exporter and non-exporter samples (e.g., -0.00547*** (0.00125) non-exporters; -0.00626*** (0.00117) exporters).
    - Tobin’s Q: positive and significant in both samples (e.g., 0.0165*** (0.00141) non-exporters; 0.00331*** (0.00104) exporters).
    - Cash: positive coefficients for both samples (e.g., 0.0605** (0.0236) non-exporters; 0.120** (0.0476) exporters).
  - Observations: 13,528 (non-exporters) and 4,621 (exporters) in each relevant column.
  - R-squared: 0.157 (non-exporter) and 0.397–0.398 (exporter).

*Source: wpiea2019093 - APPENDIX II. TABLES (PDF).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019093.pdf_