## Section 1 — Introduction and Methodology

## Source details

**Canonical URL:** [Section 1 — Introduction and Methodology](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021003-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021003-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021003-print-pdf.pdf.json)

---

### Key findings on transmission and impacts
- External financial shocks tend to have a larger effect than real shocks in driving investment and GDP growth for Asian EMs; this result is stronger at the left tails (5th percentile tail).  
- Capital flows (and their impact on corporate balance sheets) form the dominant transmission channel for EMs; less so for AEs.  
- Credit conditions are an important channel for both EMs and AEs.  
- The real exchange rate matters more in EMs than AEs, but its effect is smaller once other channels are controlled for.  
- External financial shocks have an additional impact beyond the usual channels, especially at the left tail, highlighting the role of expectations.

### Policy-relevant findings
- Foreign exchange intervention (FXI) may help dampen the capital flows and real exchange rate channels for Asian EMs and can mitigate adverse external financial shocks in the short run; its effectiveness relative to monetary policy for real shocks appears limited.  
- Macroprudential measures (MPMs) are able to mitigate the credit growth channel in both AEs and EMs more effectively than the policy rate and FXI.  
- Monetary policy transmission tends to be relatively weak in Asian EMs.

### Context, sample, shocks, and approach
- Sample: fourteen major Asian economies — 7 advanced economies (AEs) and 7 emerging markets (EMs).  
  - AEs: Australia, Hong Kong SAR, Japan, New Zealand, Singapore, South Korea and Taiwan Province of China.  
  - EMs: China, India, Indonesia, Malaysia, Philippines, Thailand and Vietnam.  
- Frequency and sample period: quarterly data from 1995q1 through 2019q2.  
- External real shocks: changes in terms of trade; a US total factor productivity (TFP) shock (Fernald 2012-19); a demand shock proxied by US defense spending (Ramey 2016) (external demand shock available only until 2015 and used as robustness check).  
- External financial shocks: an exogenous global monetary policy shock (residual of the one-year government bond rate regressed on 30-minute changes of one-month ahead Fed funds futures around FOMC announcements); and an excess bond premium (residual component of corporate bond spreads).  
- Transmission channels examined: capital flows (with disaggregation to corporates and banks), the real exchange rate, and domestic credit growth.  
- Empirical approach: panel quantile regressions with local projection methods to estimate impacts on investment and GDP growth at the median and left tail (5th percentile), using a two-step procedure for panel quantile regressions (following Canay 2011 and IMF 2019b).  
- Policy measures analyzed: FX purchases (FXI), policy interest rate changes, and macroprudential measures (MPMs). Policy exogeneity addressed via residuals from regressions of the policy on macroeconomic fundamentals and financial variables.

---

### Section 2 — Data, Methodology, and Core Results

### Data sources and variables
- GDP and investment growth: quarter on quarter change in real output and gross fixed capital formation from IMF’s WEO database.  
- Capital flows (aggregate excluding FDI, other flows to banks, other flows to corporates): IMF’s Balance of Payments Statistics Database, expressed as a share of GDP.  
- Real exchange rate: quarter on quarter change in the CPI-based REER from the IMF Information Notice System database.  
- Credit: quarter percentage change in credit from other depository corporations to the private sector from IMF’s Monetary and Financial Statistics database.  
- Policy measures:  
  - FXI: change in official FX reserves less valuation changes expressed as a share of GDP.  
  - MPMs: aggregate measure of changes in macroprudential measures using the dataset of Alam et al. (2019).  
  - Monetary policy interest rate: taken from WEO.  
- External shocks: Excess bond premium (Gilchrist and Zakrajsek, 2012); US monetary policy shock; utilization-adjusted TFP shock (Fernald, 2014); terms of trade shock from IMF WEO database.

### What matters for investment and GDP growth — summary findings
- Both real shocks (terms of trade, US TFP) and financial shocks (US monetary policy changes and excess credit spreads) significantly affect real GDP and investment growth in Asian EMs.  
- Financial shocks are the more important driver, particularly at the tails where real shocks are not statistically significant.  
- External financial conditions have a smaller and less persistent impact on AEs.  
- Shocks remain significant even when including transmission channels in regressions.

### Transmission channels — relative importance and dynamics
- Capital flows channel:  
  - Dominant transmission channel for EMs; generally not significant for AEs.  
  - For GDP growth in EMs, the balance sheet channel (non-FDI capital flows) is significant and negative in the medium term: higher capital flows today lead to lower growth 4-6 quarters out (surges tend to reverse).  
  - Capital inflows are positive for both EM and AE investment; persistent for AEs, but benefits for EMs weaken over time.  
- Credit channel:  
  - Significant for growth in both EMs and AEs but operates differently: for EMs, it boosts growth temporarily; for AEs, it acts as a shock amplifier with lower growth in the medium term.  
- Exchange rate channel:  
  - Significant for EM growth in the short-term but not for AEs. For investment, exchange rate channel is significant only for EMs and in the medium term.  
- Tail effects and persistence:  
  - Impact of transmission channels is magnified at the tail. Investment effects are generally more persistent than growth effects. External financial shocks have a direct expectation/sentiment effect beyond main transmission channels, particularly at the tails.
- Financial shocks specifics:  
  - Excess bond premium: significant for both AEs and EMs for investment and growth; impact on growth is short lived (dampens growth for the first three quarters then tends to reverse after about five quarters); more persistent effect on investment.  
  - Exogenous US monetary policy shock: significant for EMs’ GDP growth only.
- Real shocks specifics:  
  - Terms of trade shocks: significant for medium term-growth in EMs with a negative sign.  
  - US TFP shock: short-term negative effect on EMs; beneficial in the medium term for AEs; for investment, US TFP shocks have a more persistent negative effect and are not significant for AEs.
- Banking vs corporate balance sheet channels:  
  - Corporate flows significant for EM investment, not for AEs.  
  - Banking flows do not impact investment beyond the domestic credit channel.  
  - For EM GDP growth, both corporate and banking flows act as shock amplifiers in the medium term, with corporate flows relatively larger.

---

### Section 3 — Policies, Channel Decompositions, Robustness, and Implications

### Policy interactions and effectiveness (8-quarter cumulative convention)
- Estimation computes impact of a 1 standard deviation shock on investment and growth over an 8-period horizon, controlling for policies.  
- Convention for illustrative figures: blue = no policy, red = FXI (FX sales for negative shocks), gray = monetary policy (interest rate loosening for negative shocks), yellow = MPMs (MPM loosening for negative shocks). Policies measured as FX sales, a loosening in the policy interest rate, and a loosening in macroprudential measures for negative shocks.

### FX intervention (FXI)
- FXI dampens the real exchange rate channel for EMs in the short run; effect less discernable for AEs.  
- FXI purchases can reduce the positive impact of capital outflows on growth in the very near term for EMs while other policies can amplify the impact.  
- FXI is less effective for real shocks (e.g., terms of trade) but more effective against adverse financial shocks.  
- FXI can boost near-term growth under ultra-accommodative global conditions but may exacerbate long-term negative impact in a “low for long” environment proxied by the excess bond premium.

### Macroprudential measures (MPMs)
- MPMs appear effective in mitigating the effect of negative credit growth on growth in EMs and AEs.  
- For AEs, MPMs target credit better than the policy interest rate and FXI, resulting in larger and more stable investment.  
- For EMs, MPMs are less effective given shallower financial markets and potentially weaker transmission.

### Monetary policy interest rate
- Monetary policy is a broader tool and may be too blunt to address financial stability risks in AEs that are often inflation targeters; yields lower macroeconomic outturns relative to targeted MPMs.  
- Monetary policy tends to be relatively weak in Asian EMs, possibly due to low depth of the financial system, influence of advanced-economy interest rates (global financial cycle), and proximity to effective lower bounds.

### Relative effectiveness by shock type
- Effectiveness varies by whether the economy faces financial versus real shocks; FXI more effective against financial shocks amplified by unhedged FX liabilities in corporate and bank balance sheets.  
- Interaction terms provide evidence that policies can mitigate the contribution of external shocks (negative coefficient for policy×shock interaction) but results depend on specification and tail focus.

### Robustness, alternative specifications, and caveats
- Results are robust to alternate specifications and shocks (including real demand shocks).  
- Current specifications do not account for second round effects of channels on each other; multiple-equation specifications (e.g., VAR) could better explore intermediary relationships.  
- Alternate specifications: excluding interaction terms with shocks yields no significant or persistent dampening effects at the tail for EMs; including all three policies and their interaction terms together did not yield significant results for even the noninteraction terms.  
- Policy interaction section defines the external financial shock as a loosening in financial conditions (-1*e_mp) for interpretability.

### Channel-only specification and corporate FX debt amplification
- Channel-only specification (predict transmission channels from shocks, then regress outcomes on predicted channels) findings:  
  - For EMs, capital flows channel has a larger effect on investment and output than the bank channel.  
  - For AEs, both capital flows and bank channels are small.  
  - Credit growth is a key driver in both EMs and AEs: an initial boost to investment and growth is followed by a decline over the medium term.  
  - Real exchange rate appreciation has a more persistent cumulative negative effect for EMs than for AEs.
- Corporate FX debt robustness: interaction of shocks with corporate FX debt (share of GDP) is significant and negative primarily for financial shocks (particularly the excess credit spread), indicating high corporate FX debt amplifies adverse financial shocks for output and investment.

### Selected empirical highlights (reported regressions, observations = 370)
- Investment/GDP (median): credit growth coefficient example: 1.268*** (p-value shown as 0).  
- Investment/GDP (5th percentile): US TFP example: -0.692** (p = 0.0239); constant term example: -4.674*** (p = 0).  
- GDP growth (median): REER example: 0.363*** (p = 0.00931); credit growth example: 0.303*** (p = 0.00910); GZ shock example: -0.546*** (p = 0.00111).  
- GDP growth (5th percentile): capital flows ex FDI example: -9.031** (p = 0.0239); MP shock example: -1.096*** (p = 0.000660).

### Policy implications and concluding remarks
- The global financial cycle is central to macrofinancial fluctuations in Asian economies; external financial shocks matter substantially for small open economies as trade and financial integration advance.  
- For EM Asia: capital flows and domestic financial conditions drive output and investment; corporate balance sheets (notably FX exposure) are a source of vulnerability; expectations and cumulative accommodative global financial conditions (“low for long”) affect output and investment beyond standard channels.  
- For AEs: credit is a key channel affecting output and investment.  
- Policy takeaway: multiple policy instruments can mitigate adverse impacts of external shocks; evidence supports using different instruments under different circumstances given structural characteristics such as balance sheet mismatches, low financial development, and weak monetary policy transmission.  
- The analysis does not identify optimal policy choices; calibrated structural models and multiple-equation specifications would be useful for determining optimal responses and fully capturing credibility and medium-term inflation expectation effects.

*Prepared by Pragyan Deb, Sanaa Nadeem, and Shanaka J. Peiris; Section 1–3 of WP/21/3 — The Transmission of External Shocks in Asia: Country Characteristics and Policy Responses (January 2021).*

### Section 1

### Section 1 — Introduction and Methodology

### Key findings on transmission and impacts
- External financial shocks tend to have a larger effect than real shocks in driving investment and GDP growth for Asian EMs; this result is stronger at the left tails (5th percentile tail).  
- Capital flows (and their impact on corporate balance sheets) form the dominant transmission channel for EMs; less so for AEs.  
- Credit conditions are an important channel for both EMs and AEs.  
- The real exchange rate matters more in EMs than AEs, but its effect is smaller once other channels are controlled for.  
- External financial shocks have an additional impact beyond the usual channels, especially at the left tail, highlighting the role of expectations.

### Policy-relevant findings
- Foreign exchange intervention (FXI) may help dampen the capital flows and real exchange rate channels for Asian EMs and can mitigate adverse external financial shocks in the short run; its effectiveness relative to monetary policy for real shocks appears limited.  
- Macroprudential measures (MPMs) are able to mitigate the credit growth channel in both AEs and EMs more effectively than the policy rate and FXI, supporting the assignment of well-targeted macroprudential tools for financial stability risks.  
- Monetary policy transmission tends to be relatively weak in Asian EMs.

### Context and motivation
- Asian economies are increasingly integrated to the global economy through trade and financial linkages, exposing them to the international financial cycle. Since the global financial crisis, prolonged accommodative financial conditions in advanced economies have fueled search-for-yield behavior and large, volatile capital flows in Asian EMs.  
- Structural vulnerabilities identified for Asian EMs include shallow financial markets, large and unhedged FX liabilities in corporates and banks, and in some cases weak monetary policy transmission.  
- The paper focuses on three transmission channels: (i) corporate balance sheet effects (proxied by non-FDI capital flows); (ii) the competitiveness channel (proxied by the real exchange rate); and (iii) credit conditions (growth in credit to the private sector).

### Data, sample, and shocks considered
- Sample: fourteen major Asian economies — 7 advanced economies (AEs) and 7 emerging markets (EMs).  
  - AEs: Australia, Hong Kong SAR, Japan, New Zealand, Singapore, South Korea and Taiwan Province of China.  
  - EMs: China, India, Indonesia, Malaysia, Philippines, Thailand and Vietnam.  
- Frequency and sample period: quarterly data from 1995q1 through 2019q2 (with some country/specification limitations due to data availability).  
- External real shocks: changes in terms of trade; a US total factor productivity (TFP) shock (Fernald 2012-19); a demand shock proxied by US defense spending (Ramey 2016). (The external demand shock variable is available only until 2015 and is used as a robustness check rather than baseline.)  
- External financial shocks: an exogenous global monetary policy shock (estimated using the residual of the one-year government bond rate regressed on 30-minute changes of the one-month ahead Fed funds futures around FOMC announcements); and an excess bond premium (residual component of corporate bond spreads not explained by expected default risk, estimated on a sample of US nonfinancial firms).  
- Transmission channels examined: capital flows (with disaggregation to corporates and banks), the real exchange rate, and domestic credit growth. A domestic financial conditions index was considered but credit growth provided clearer identification.

### Empirical approach and specifications
- Estimation strategy: panel quantile regressions with local projection methods to estimate impacts on investment and GDP growth at the median and left tail (5th percentile). A two-step procedure for panel quantile regressions is used (following Canay 2011 and IMF 2019b):  
  - Step 1: estimate unobserved fixed effects using within-estimators.  
  - Step 2: adjust the dependent variable with fixed effects and estimate standardized conditional quantile regressions.  
- Basic conditional equation for h-quarter-ahead impacts at quantile q: specification (1) with ∆^h dep̃_i,t+h,q = α_h,q + β_h,q shocks_i,t + γ_h,q X_t−1 + ε_i,t,h,q, where β_h,q measures the direct effect of external shocks.  
- Transmission-channel regressions: specification (2) estimates the impact of external shocks on each transmission channel (real exchange rate, capital flows, domestic credit) at horizon h.  
- Full-channel decomposition: specification (3) assesses the impact of shocks through transmission channels on GDP growth or investment, with θ_h,q measuring the effect of the external shock outside the transmission channels. For capital flows the analysis disaggregates flows to corporates and banks; transmission channel variables are purged of the impact of global shocks by first-stage regression (the components orthogonal to shocks are used). Standard errors are obtained via bootstrapping to account for estimated dependent variables.  
- Policy interactions: specification (4) adds interaction terms between policies and transmission channels and shocks to evaluate policy effectiveness: ∆^h dep̃_i,t+h,q = α_h,q + β_h,q transmission_i,t + θ_h,q shocks_t + δ_h,q policy_i,t + η_h,q (transmission_i,t ∙ policy_i,t) + μ_h,q (shocks_i,t ∙ policy_i,t) + γ_h,q X_t−1 + ε_i,t,h,q. The net effect of a policy on a transmission channel is β_h + η_h (and θ_h + μ_h for shocks). Policies are introduced one at a time in each specification.  
- Policy measures: FX purchases (FXI), a tightening in the policy interest rate, and tightening in macroprudential measures (MPMs). To address endogeneity, the exogenous component of each policy is identified via residuals from regressions of the policy on macroeconomic fundamentals and financial variables (following Brandao Marques et al. 2020 and IMF 2020).

### Contribution and relation to literature
- Extends IMF 2019a by analyzing drivers of the real economy in addition to exchange rate and financial conditions, comparing relative contributions of channels, and introducing the role of policies.  
- Confirms and extends IMF 2020 results for Asian EMs by examining dynamic responses (not only contemporaneous effects) and additional shocks and impact variables.  
- Consistent with Brandao Marques et al. 2020 on the role of MPMs in dampening the credit channel; also finds evidence supporting a role for FXI in emerging Asia to mitigate volatility in capital flows, the exchange rate, and external financial shocks in the short run.

*Prepared by Pragyan Deb, Sanaa Nadeem, and Shanaka J. Peiris; Section 1 of WP/21/3 — The Transmission of External Shocks in Asia: Country Characteristics and Policy Responses (January 2021).*

### Section 2

### Section 2

### Data and methodology
- Panel dataset with a quarterly frequency on 14 Asian economies.
- GDP and investment growth: quarter on quarter change in real output and gross fixed capital formation from IMF’s World Economic Outlook (WEO) database.
- Capital flows series (aggregate excluding FDI, other flows to banks, other flows to corporates): IMF’s Balance of Payments Statistics Database, expressed as a share of GDP.
- Real exchange rate: quarter on quarter change in the CPI-based REER from the IMF Information Notice System database.
- Credit: quarter percentage change in credit from other depository corporations to the private sector from IMF’s Monetary and Financial Statistics database.
- Policy measures:
  - FXI: change in official FX reserves less valuation changes expressed as a share of GDP (following IMF 2019a).
  - MPMs: aggregate measure of changes in macroprudential measures (tightening or loosening) using the dataset of Alam et al. (2019).
  - Monetary policy interest rate: taken from WEO.
- External shocks:
  - Excess bond premium (Gilchrist and Zakrajsek, 2012).
  - US monetary policy shock (as in Albrizio et al. forthcoming).
  - Utilization-adjusted TFP shock (Fernald, 2014).
  - Terms of trade shock from IMF WEO database, expressed as a quarter on quarter change.
- Estimation strategy:
  - Regressions of output and investment on exogenous shocks and transmission channels.
  - Compute impact of a one standard deviation shock to each regressor over an 8-quarters horizon.
  - Interaction terms between policies and shocks included to capture dampening/amplifying effects.

### What matters for investment and GDP growth — key findings
- Both real shocks (terms of trade, US TFP) and financial shocks (US monetary policy changes and excess credit spreads) significantly affect real GDP and investment growth in Asian EMs.
- Financial shocks are the more important driver, particularly at the tails where real shocks are not statistically significant.
- External financial conditions have a smaller and less persistent impact on AEs.
- Shocks remain significant even when including transmission channels in regressions.

### Transmission channels — relative importance and dynamics
- Capital flows channel:
  - Dominant transmission channel for EMs; generally not significant for AEs.
  - For GDP growth in EMs, the balance sheet channel (proxied by non-FDI capital flows) is significant and negative in the medium term: higher capital flows today lead to lower growth 4-6 quarters out (suggesting surges tend to reverse).
  - Capital inflows are positive for both EM and AE investment; persistent for AEs, but benefits for EMs weaken over time.
- Credit channel:
  - Significant for growth in both EMs and AEs but operates differently:
    - For EMs, it boosts growth temporarily.
    - For AEs, it acts as a shock amplifier with lower growth in the medium term.
- Exchange rate channel:
  - Significant for EM growth in the short-term but not for AEs.
  - For investment, exchange rate channel is significant only for EMs and in the medium term.
- Tail effects and persistence:
  - Impact of transmission channels is magnified at the tail.
  - For EM growth, capital flows and terms of trade shocks generate large volatilities in the medium term; global monetary and financial shocks generate a short term drag.
  - Investment effects are generally more persistent than growth effects.
  - External financial shocks have a direct expectation/sentiment effect beyond main transmission channels, particularly at the tails.
- Financial shocks specifics:
  - Excess bond premium significant for both AEs and EMs for both investment and growth; impact on growth is short lived and shows mean reversion—dampens growth significantly for the first three quarters then tends to reverse after about five quarters; more persistent effect on investment.
  - Exogenous US monetary policy shock is only significant for EMs’ GDP growth.
- Real shocks:
  - Terms of trade shocks significant for medium term-growth in EMs with a negative sign (gains are short lived and associated with competitiveness loss in the medium term).
  - US TFP shock: short-term negative effect on EMs; beneficial in the medium term for AEs. For investment, US TFP shocks have a more persistent negative effect and are not significant for AEs.
- Banking vs corporate balance sheet channels:
  - Corporate flows are a significant explanator for EM investment, but do not matter in AEs.
  - Banking flows do not have an impact on investment in both EMs and AEs beyond their impact via the domestic credit channel.
  - For GDP growth in EMs, both corporate and banking flows act as shock amplifiers in the medium term, with corporate flows playing a relatively larger role.
  - The relative importance of corporate vs. bank balance sheets in emerging Asia could reflect that corporates are generally less hedged than banks.

### The role of policies — effectiveness and interactions with shocks
- Estimation of policy interactions via equation (4) computes impact of a 1 standard deviation shock on investment and growth over an 8-period horizon, controlling for policies.
- General result: capital flows and credit continue to drive investment and GDP growth for EMs; credit channel and external financial shocks (excess bond premium) are main explanators in AEs. Interaction terms provide evidence that policies can mitigate the contribution of external shocks (negative coefficient for policy×shock interaction).
- Illustrative convention used in figures:
  - Blue lines: cumulative impact over eight quarters of a one standard deviation change without policies (β̂ coefficient).
  - Red lines: impact with FXI (FX sales in response to negative shocks).
  - Gray lines: specification with monetary policy (interest rate loosening for negative shocks).
  - Yellow lines: specification with MPMs (MPM loosening for negative shocks).
  - Policy responses measured as FX sales, a loosening in the policy interest rate, and a loosening in macroprudential measures for negative shocks.
- FXI (foreign exchange intervention):
  - Dampens the real exchange rate channel for EMs in the short run; effect less discernable for AEs.
  - Larger effect when re-estimating keeping only negative shocks; FXI purchases can reduce the positive impact of capital outflows on growth in the very near term for EMs while other policies can amplify the impact.
  - Positive effect on GDP growth for EMs is more short-lived.
  - FXI is less effective for real shocks (e.g., terms of trade) but more effective against adverse financial shocks.
- Macroprudential measures (MPMs):
  - Appear effective in mitigating the effect of negative credit growth on growth in EMs and AEs.
  - For AEs, MPMs target credit better than the policy interest rate and FXI, resulting in larger and more stable investment.
  - For EMs, MPMs are less effective given shallower financial markets and potentially weaker transmission.
- Monetary policy interest rate:
  - Policy interest rate is a broader tool and may be too blunt to address financial stability risks in AEs that are often inflation targeters; yields lower macroeconomic outturns relative to targeted MPMs.
- Relative effectiveness correlates with shock type:
  - Policies’ effectiveness varies by whether the economy faces financial versus real shocks; FXI more effective against financial shocks amplified by unhedged FX liabilities in corporate and bank balance sheets.

### Caveats and robustness
- Results are robust to alternate specifications and shocks (including real demand shocks).
- Current specifications do not account for second round effects of the channels and shocks on each other; a multiple equation specification, as in a vector autoregression (VAR) setting, could better explore intermediary relationships.
- Alternate specifications tested:
  - Excluding interaction terms with shocks: for EMs, no significant or persistent dampening effects of any of the policies at the tail found in that exercise.
  - Alternate specification with all three policies and their interaction terms did not yield significant results for even the noninteraction terms.
- Definitions and interpretation note:
  - For the policy interaction section, the external financial shock is defined as a loosening in financial conditions (-1*e_mp) for ease of interpretation; with variables expected to have a positive β̂ coefficient, a dampening (amplifying) interaction term would have a negative (positive) coefficient.

*Source: wpiea2021003-print-pdf - Section 2*

### Section 3

### Impact of policies on real vs. financial shocks, EM investment, median

### Cumulative shock framework and key mechanism notes
- Figures analyze the cumulative impact over 8 quarters of a one standard deviation capital outflow with a one standard deviation policy response (FX sales, interest rate loosening, or MPM loosening).
- FX intervention (FXI) dynamics:
  - FXI can boost near-term growth under ultra-accommodative financial conditions but may exacerbate long-term negative impact on growth in a “low for long” environment, proxied by the excess bond premium.
  - In a specification with the excess bond premium, a spike (drop) in excess bond premium lowers (raises) growth in the first 3-4 quarters, followed by an increase (decline) in growth.
  - Use of FXI could lower growth in the medium term for EMs more than the baseline scenario; the cost from FXI for AEs appears not as large, possibly due to deeper financial markets.
- Monetary policy in EMs:
  - Monetary policy is less effective than FXI (for financial shocks) and less effective than macroprudential measures (MPMs) for the credit channel.
  - Potential reasons: low depth of the financial system, influence of advanced-economy interest rates (global financial cycle), and proximity to effective lower bounds for interest rates.

### Robustness checks and transmission-channel–only specification
- Alternate specification: impact on investment and GDP growth of only the transmission channels, excluding the direct impact of external shocks, using predicted values from regressions of individual channels on external shocks (estimated equation (3) in the text).
  - Equation (3) form:
    ∆_h dep̃_i,t+h,q = α_h,q ∆_h dep̃_i,t+h−1,q + β_h,q transmission̂_i,t + ε_i,t,h,q
- Transmission channels considered: real exchange rate, credit growth, balance sheets via capital flows (aggregate and decomposed between banks and corporates).
- Key findings from channel-only specification:
  - For EMs, the capital flows channel has a relatively larger effect on investment and output than the bank channel. Interpretation: banks may have greater access to FX hedging or regulatory limits on FX exposure compared with corporates.
  - For AEs, both capital flows and bank channels are small.
  - Credit growth is a key driver in both EMs and AEs: an initial boost to investment and growth is followed by a decline over the medium term, highlighting medium-term financial stability risks.
  - Real exchange rate appreciation has a more persistent cumulative negative effect for EMs than for AEs, suggesting a stronger competitiveness channel in EMs.

### Corporate FX debt as an amplification channel
- Robustness check focusing on corporate balance sheets:
  - Replace capital flows with FX debt of nonfinancial corporates (share of GDP) and interact shocks with corporate FX debt.
  - Interaction terms of corporate FX debt are significant and negative primarily for financial shocks (particularly the excess credit spread) rather than for real shocks.
  - Interpretation: in the event of an adverse financial shock, high FX debt in corporates amplifies negative effects on output and investment growth.
  - Result holds for both AEs and EMs.

### Empirical table highlights (selected)
- Tables show median and 5th percentile impacts on Investment/GDP and GDP growth across multiple specifications (observations = 370 in displayed regressions).
- Selected statistically significant coefficients (p-values in parentheses; significance codes in tables: *** p<0.01, ** p<0.05, * p<0.1):
  - Investment/GDP (median): Credit growth coefficients frequently positive and highly significant, e.g., 1.268*** (p-value shown as 0) in one specification.
  - Investment/GDP (5th percentile): US TFP often negative and significant, e.g., -0.692** (p = 0.0239) in one specification; constant terms large and negative in 5th percentile regressions (e.g., -4.674*** with p = 0).
  - GDP growth (median): REER positive and significant in some specifications, e.g., 0.363*** (p = 0.00931); credit growth positive and significant, e.g., 0.303*** (p = 0.00910); GZ shock negative and significant, e.g., -0.546*** (p = 0.00111).
  - GDP growth (5th percentile): Capital flows ex FDI sometimes large negative and significant (e.g., -9.031** with p = 0.0239 in one specification); MP shock coefficients vary in sign and significance across specifications (e.g., -1.096*** with p = 0.000660 in one specification).
- Observations across reported regressions: 370.

### Policy implications and concluding remarks
- The global financial cycle is central to macrofinancial fluctuations in Asian economies; as trade and financial integration advance, external financial shocks matter substantially for small open economies.
- For EM Asia:
  - Capital flows and domestic financial conditions drive output and investment.
  - Corporate balance sheets (notably FX exposure) are a source of vulnerability, potentially reflecting lower FX hedging relative to banks.
  - Expectations and cumulative accommodative global financial conditions (“low for long”) affect output and investment beyond standard transmission channels.
- For AEs:
  - Credit is a key channel affecting output and investment.
- Policy takeaways:
  - Multiple policy instruments can mitigate adverse impacts of external shocks; evidence shows merits to using one policy versus another under certain circumstances, providing a rationale for Asian policymakers’ use of multiple instruments.
  - Structural characteristics that strengthen the case for multiple instruments include balance sheet mismatches, low financial development, and weak monetary policy transmission.
  - The analysis does not identify optimal policy choices; calibrated structural models and multiple-equation specifications would be useful for determining optimal responses and fully capturing credibility and medium-term inflation expectation effects.
- Suggested further work:
  - Broaden analysis with multiple-equation specifications.
  - Use structural models to parse optimal policy choices for specific shocks.
  - More closely examine banking and corporate channels in emerging Asia given their large FX asset and liability positions and potential amplification role.

*Source: wpiea2021003-print-pdf - Section 3*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021003-print-pdf.pdf_
