## 1. Summary: Determinants of the Sensitivity of Domestic FCIs

## Source details

**Canonical URL:** [1. Summary: Determinants of the Sensitivity of Domestic FCIs](https://www.imf.org/-/media/files/publications/wp/2018/wp1815.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp1815.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp1815.pdf.json)

---

### Major findings
- A single factor, “global financial conditions,” accounts for a large share of variation in domestic financial conditions globally and moves in tandem with the U.S. FCI and measures of global risk such as the VIX.
- No conclusive evidence that the global factor has gained significant influence over the past two decades; rolling-window statistics display cyclical patterns but a broadly flat trajectory.
- About 20 to 40 percent of the variation in domestic FCIs across countries can be attributed to global financial conditions; panel regressions yield about 30 percent on average while some economies reach almost 70 percent.
- Domestic monetary policy shocks account for about 15 percent of the variation across countries with flexible exchange rates (panel VAR average).
- Local FCIs react more rapidly and strongly to global financial shocks than to changes in domestic policy rates.
- Greater financial development reduces sensitivity of domestic FCIs to global financial shocks.
- Financial linkages with the United States (especially FDI stock) are the most reliable indicator of higher synchronization with global financial conditions; portfolio and banking linkages show mixed or insignificant signs.
- The three-factor dynamic factor model explains greater than 40 percent of FCI variability on average; a one-factor PCA-based global measure closely tracks the DFM global factor.

### Construction and measurement of FCIs
- FCIs estimated for 1990–2016 at monthly frequency for 43 advanced and emerging market economies.
- Estimation method: time-varying factor augmented vector autoregressive model (TVP-FAVAR).
- FCIs are purged of current macroeconomic conditions and exclude exchange rate to avoid overstating global influence.
- Base FCI variables include: corporate spreads, term spreads, interbank spreads, sovereign spreads, changes in long-term interest rates, equity returns, house price returns, equity return volatility, changes in the market share of the financial sector, and credit growth.

### Global factor and robustness
- The average correlation between the U.S. FCI and two global measures and the VIX is 82 percent.
- Rolling 36- and 60-month windows, Forbes-Rigobon adjustments, and regional robustness checks support the lack of a sharp uptrend in the variance explained by global financial conditions.
- PCA one-factor and three-factor DFM representations produce consistent identification of a global factor and regional factors (“Emerging market”, “Euro area”, “Global financial crisis”).

### Determinants of sensitivity to global financial conditions (panel regression results)
- Baseline panel regression: FCI_c,t = α_i + β1 FCI_t^US + β2 FCI_c,t−1 + β3 FCI_t^US * FCI_c,t−1 + β4 CCHAR_c,t−1 + ε_it
- Key estimated signs and significance:
  - Direct effect of U.S. FCI: Estimated sign +; Significance: ***
  - FDI linkages with the U.S.: Estimated sign +; Significance: **
  - Trade openness: Estimated sign +; Significance: **
  - Financial development: Estimated sign −; Significance: **
  - Portfolio linkages with the U.S.: Estimated sign −; Not significant
  - Banking linkages with the U.S.: Estimated sign −; Not significant
  - Exchange rate flexibility: Estimated sign + (expected −)
- Interpretation:
  - Stronger FDI and financial openness increase sensitivity.
  - Deeper financial development and market depth attenuate global shock transmission.
  - No clear robust role for exchange rate regime or general capital account openness in baseline specifications.

### Transmission and VAR evidence (panel VAR, 2001–2016; 25 small open economies with flexible exchange rates)
- Baseline VAR variables: Y_c,t = [FCI_t^* , ∆IP_c,t , ∆CPI_c,t , FCI_c,t , Δi_c,t ]' with FCI_t^* = U.S. FCI.
- Estimation: Four lags; Pesaran, Shin, and Smith (1999) mean group estimator; Cholesky decomposition ordering U.S. FCI first, domestic interest rates last.
- Impulse responses:
  - A one standard deviation tightening in global financial conditions (U.S. FCI) raises domestic financial conditions by, on average, about one-fifth of a standard deviation across the sample.
  - Local FCIs react faster and more strongly to global financial shocks than to domestic policy rate changes, implying monetary policy must react quickly and strongly to offset global shocks.
- Peak impulse response robustness (peak magnitudes and times to peak, In percent and Months):
  - Example rows:
    - Exog: commodity price inflation — Monetary Policy Shock: Peak Magnitude 16.5; Time to Peak 8; Global Financial Shock: Peak Magnitude 15.6; Time to Peak 5
    - Endog: change in NEER — Monetary Policy Shock: Peak Magnitude 15.4; Time to Peak 9; Global Financial Shock: Peak Magnitude 20.1; Time to Peak 5
    - Endog: change in bilateral exchange rate with the U.S. — Monetary Policy Shock: Peak Magnitude 15.4; Time to Peak 9; Global Financial Shock: Peak Magnitude 20.5; Time to Peak 5

### Variance decompositions (shares attributable to shocks)
- Panel VAR average (small open economies with flexible exchange rates):
  - Global financial shocks: about 21 percent of variation in domestic FCIs.
  - Domestic monetary policy shocks: about 15 percent of fluctuations in FCIs.
- Individual-country VAR averages:
  - Global financial conditions: about 40 percent of domestic FCI variations.
  - Monetary policy shocks: about 12 percent of domestic FCI variations.
- Table 6 (variance decomposition at a 12-month horizon; In percent) examples:
  - Exog: commodity price inflation — Global Financial Conditions 15.6; Domestic Monetary Policy 15.5; Other 68.9
  - Exog: commodity price inflation, global IP growth, U.S. shadow rate (Xia Wu) — Global 16.8; Domestic 11.8; Other 71.4
  - Endog: change in REER — Global 15.1; Domestic 12.6; Other 72.3
- Individual-country VAR summaries (Table 7; In percent) examples:
  - Small Open Advanced Economies, Baseline: World FCI 26.1; Domestic Policy Rate 20.1; Domestic FCI 28.4
  - Emerging Markets, Baseline: World FCI 38.2; Domestic Policy Rate 11.3; Domestic FCI 25.6
- Lag sensitivity and specification robustness reported in tables; results broadly robust across lags and specifications.

### Country heterogeneity and identification refinements
- Country-by-country VARs show considerable variation; emerging market economies often exhibit higher shares attributable to global financial conditions, sometimes exceeding 60 percent.
- Distribution of shares can be bi-modal with small open advanced economies clustering at lower shares.
- More refined monetary policy identification (Gertler and Karadi (2015) three-step, high-frequency surprises instrument) applied for Australia, New Zealand, Norway, Sweden yields impulse responses similar to Cholesky-based responses.
- For the four case studies, average shares: global financial conditions 15 percent; domestic monetary policy 33 percent (Cholesky decomposition).

### Temporal stability
- Comparison between 2001–07 and 2010–16 shows no discernible change in the importance of global financial conditions; variance decompositions broadly stable and not statistically different at the 95 percent level.

### Policy implications and recommendations
- Monitor and construct FCIs purged of macroeconomic fundamentals to distinguish non-fundamental spillovers from fundamentals-driven comovements.
- Strengthen financial development to reduce sensitivity to global financial shocks.
- Recognize that greater financial linkages increase exposure to global cycles; incorporate external factors into domestic policy setting.
- Policymakers should prioritize speed and preemptive elements in policy design given the faster reaction of local financial conditions to global shocks.
- Prudential tools may face leakage and limited efficacy in open economies; coordination and timely response frameworks may be required.

### Australia: country-specific summary and implications
- Findings for Australia align with cross-country results:
  - Global financial conditions capture a significant share of domestic FCI dynamics; domestic FCIs respond faster to global shocks than to domestic policy changes.
  - Panel VAR decompositions for precrisis (2001–07) and postcrisis (2010–16) samples show shares and confidence bands reported for Australia among flexible-exchange-rate countries.
- Policy implications for Australia:
  - Account for rapid transmission of global financial shocks in response strategies.
  - Financial development can act as a buffer and remain a policy priority to enhance resilience.
- Suggested research avenues for Australia:
  - Examine asymmetries and threshold effects in responses to tightening versus loosening global conditions.
  - Assess the relative influence of temporary acute global tightening episodes versus prolonged accommodative periods.
  - Further explore cross-country transmission channels and the role of different financial intermediaries.

*Source: wp1815 - 1. Summary: Determinants of the Sensitivity of Domestic FCIs (authors’ estimates and figures as provided).*

### 1. Summary: Determinants of the Sensitivity of Domestic FCIs ............................................21

### 1. Summary: Determinants of the Sensitivity of Domestic FCIs ............................................21

### Major analytic sections
- Determinants of Sensitivity to Global Financial Conditions ...............................................23
- Robustness: Determinants of Sensitivity to Global Financial Conditions ...........................24
- Summary: Panel VAR Robustness: Impulse Response of Domestic FCI to .......................27
- Summary: Individual Country VAR Robustness Impulse Response ..................................28
- Summary Panel VAR Robustness: Share of Domestic FCI Variation ................................30
- Summary: Individual Country VAR Robustness: Share of Domestic FCI .........................31

### Figures referenced in the chapter
- United States: Financial Conditions Indices, 1991–2016 ....................................................13
- Selected Advanced and Emerging Market Economies ........................................................14
- Three-Factor Model Based on Financial Conditions Index, 1995–2016 .............................15
- Single Factor Versus Principal Component Analysis, 1995–2015 ......................................16
- Variance Accounted for by One and Three Factor Models .................................................17
- Variance Attributable to Global Conditions, 1995–2016 ....................................................18
- Robustness: Attributable to Global Financial Conditions, 1995–2016 ...............................18
- Robustness: Variance Attributable to Global Financial Conditions, ...................................19
- Response of Domestic Financial Conditions to Shocks ......................................................26
- Share of Domestic Financial Conditions Index Fluctuations Attributable to ....................29
- Share of Domestic Financial Conditions Index Fluctuations Attributable to ....................32
- Selected Advanced Economies: Response of Financial Conditions Index ........................34
- Share of Domestic Financial Index Fluctuations Attributable to Global Financial ...........35

### Appendix material associated with the chapter
- Appendix Table 1. Country Coverage ................................................................................................................36
- Appendix Table 2. Data Sources ........................................................................................................................37

*Source: wp1815 - 1. Summary: Determinants of the Sensitivity of Domestic FCIs (pages and figure listings as provided).*

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

### wp1815 - References

### I. INTRODUCTION
- Research question: To what extent can individual countries steer domestic financial conditions in a globally integrated financial system?
- Debate:
  - Rey (2016, 2013): In a world of freely flowing capital, U.S. monetary policy shocks spill over and affect domestic financial conditions; fluctuating exchange rates cannot insulate economies from the global financial cycle; independent monetary policy is possible only with restricted capital mobility.
  - Kamin (2010), Obstfeld (2015), Klein and Shambaugh (2015), Obstfeld, Ostry, and Qureshi (2017), Disyatat and Rungcharoenkitkul (2016), IMF (2017): Exchange rate flexibility allows at least some monetary autonomy; policymakers retain considerable sway over domestic financial conditions.
- Definition and role of financial conditions:
  - Financial conditions broadly reflect how easy it is to obtain financing, summarizing price and nonprice costs of credit.
  - Monetary policy seeks to influence inflation and output largely through effects on financial market variables (asset prices, bank credit volumes, collateral valuations, term premiums) along with direct effects through policy rates.
- Key conceptual points:
  - High comovement of domestic financial conditions across countries is not per se evidence of loss of policy autonomy; may reflect comovements in fundamentals.
  - Global financial integration can, however, weaken transmission channels of domestic monetary policy or diminish prudential policy impact, and expose countries to shocks unwarranted by fundamentals (e.g., investor sentiment shifts).
- Paper focus:
  - Importance of common global components underlying domestic financial conditions, their evolution over time, key drivers, country characteristics affecting sensitivity to global factors, and the ability of monetary policy to influence domestic financial conditions.
  - Development of new FCIs purged of macroeconomic factors and comparable across advanced and emerging market economies.

- Main results:
  - A single factor, “global financial conditions,” accounts for a large share of variation in domestic financial conditions globally; this factor moves in tandem with the U.S. FCIs and measures of global risk, such as the VIX.
  - No conclusive evidence that this global factor has gained significant influence over the past two decades.
  - Financial linkages (such as cross-country investments) are the most reliable indicator of global financial conditions’ influence on local FCIs.
  - Greater financial development can reduce sensitivity of domestic FCIs to global financial shocks.
  - About 20 to 40 percent of the variation in domestic FCIs across countries can be attributed to global financial conditions; domestic factors account for the rest.
  - Domestic monetary policy shocks account for about 15 percent of the variation across countries with flexible exchange rates.
  - Local financial conditions react more rapidly to global financial shocks than to changes in domestic policy rates, complicating timely policy responses.

### II. AN OVERVIEW OF FINANCIAL CONDITIONS
- A. Financial Conditions: Main Concepts and Selected Literature
  - Definitions:
    - Hatzius and others (2010): Current state of financial variables that influence economic behavior and thereby the future of the economy.
    - Carlson, Lewis, and Nelson (2012): Connected to price and nonprice costs of credit.
    - This paper: Gauges costs, conditions, and availability of domestic funds to the local economy; includes interest rates, asset price valuations, risk appetite, and willingness to hold illiquid assets.
  - Transmission channels of monetary policy through financial conditions:
    - Traditional/New Keynesian channels: policy rate changes and expectations affecting longer-term rates; exchange rate effects on trade.
    - Nontraditional channels: credit supply imperfections, balance sheet channel, bank capital channel, risk-taking channels.
  - FCIs:
    - Aim to distill information from a broad array of financial variables including measures of risk taking and financial frictions, capturing credit constraints and external financing premiums.
    - FCIs can be more helpful in predicting economic activity than indicators of current and past real economic activity (Hatzius and others 2010; Koop and Korobilis 2014; Adrian, Boyarchenko, and Giannone 2016).
  - Drivers of financial conditions besides policy:
    - Uncertainty about exposures of major financial players, shocks to borrowers’ net worth not triggered by policy, runs, changes in risk perception, investor sentiment.

- B. Transmission of Financial Conditions Across Countries
  - Mundell-Fleming trilemma: policymakers can choose only two of fixed exchange rates, free capital mobility, and monetary autonomy; flexible exchange rates can insulate from external shocks in theory.
  - In practice, financial conditions can be transmitted across countries in ways not fully offset by exchange rates (Obstfeld 2015); exchange rate movements can induce changes in domestic financial conditions and amplify external shocks (Kearns and Patel 2016).
  - Transmission mechanisms: changes in credit volumes, capital flows, comovements in risk premiums affecting collateral valuation and borrowing constraints.
  - Two ways global integration complicates management of domestic financial conditions:
    - Policymakers must weigh external factors when pursuing domestic objectives.
    - Global integration may hamper transmission of monetary policy or reduce effectiveness of prudential policies; external shocks can affect local conditions abruptly.
  - Empirical literature highlights role of global “push factors” (e.g., VIX) and a global factor or “global financial cycle” driven by U.S. monetary policy shocks (Calvo, Leiderman, and Reinhart 1996; Bruno and Shin 2013; IMF 2014a; Fratzscher 2012; Baskaya and others 2017; Miranda-Agrippino and Rey 2015; Rey 2016).
  - Distinction between fundamentals-driven comovement and non-fundamental-driven spillovers is crucial but empirically difficult; paper focuses on FCIs purged of macroeconomic fundamentals to address this issue.

- C. Constructing Financial Conditions Indices Across Advanced and Emerging Market Economies
  - Gap: Most literature and FCIs focus on a few advanced economies, especially the U.S.; FCIs for emerging market economies are rare.
  - Approach:
    - Latent FCIs extracted from an array of financial variables while taking account of growth and inflation.
    - Estimation method: time-varying factor augmented vector autoregressive model (TVP-FAVAR) of Koop and Korobilis (2014), building on Primiceri (2005) TVP-VAR and Doz, Giannone, and Reichlin (2011) dynamic factor models.
    - Model form (concept): x_t = λ_t^y Y_t + λ_t^f f_t + u_t; [Y_t; f_t] dynamics via time-varying B matrices and shocks ε_t.
  - Advantages of TVP-FAVAR:
    - Purges FCI of effects of macroeconomic conditions (initially purged of current macroeconomic conditions).
    - Time-varying parameters account for evolving relationships, policy regime changes, and varying transmission intensity of financial shocks.
  - Variable selection principles:
    - Exclude direct measures of ease of access to international markets to focus on global factors as drivers of domestic conditions.
    - Include variables consistent across countries capturing equity, housing, bond, and interbank markets: interest rates and spreads (longer-term rates, corporate, interbank, term spreads), asset price returns (equity and house price returns), equity return volatility, credit growth, and, where available, survey-based lending standards.
    - Exclude exchange rate from the FCI to avoid overstating global influence via direct inclusion.
  - Implementation details:
    - FCIs estimated for 1990–2016 at monthly frequency for 43 advanced and emerging market economies.
    - FCIs are based on 10 underlying financial indicators: corporate spreads, term spreads, interbank spreads, sovereign spreads, changes in long-term interest rates, equity returns, house price returns, equity return volatility, changes in the market share of the financial sector, and credit growth.
    - Robustness: Additional financial variables (e.g., lending standards, sovereign spreads) tested where data permit; results broadly similar.

### III. FINANCIAL CONDITIONS AROUND THE WORLD
- A. Financial Conditions Indices: Selected Countries and Stylized Facts
  - U.S. FCI:
    - U.S. FCI developed in the paper closely tracks counterparts developed by IMF and Federal Reserve Banks of Chicago and Kansas City for 1990–2016.
    - Interpretation: positive values = tighter-than-average financial conditions; negative values = looser-than-average.
    - Historical episodes captured:
      - Early 1990s tranquility; tightening around 1998 (Long-Term Capital Management), dot-com crash in 2000, accounting and corporate failures around 2002 (Arthur Andersen, WorldCom).
      - Global financial crisis in 2008 produced unprecedented spike in FCI.
      - Post-crisis gradual uptrend though conditions broadly accommodative.
  - Selected small open economies:
    - Russia: dramatic tightening in 1998, spike larger than during the global financial crisis.
    - Korea: tighter during the global financial crisis than during the Asian financial crisis (1997–98).
    - Chile: global financial crisis represents the sharpest FCI spike of the past two decades.
    - Netherlands (small open euro area economy): financial conditions tightened almost equally during the euro area crisis and the global financial crisis.
  - Underlying drivers across countries:
    - Factor loadings indicate interbank and corporate spreads, equity return volatility, and changes in house prices are the top contributors to countries’ FCIs for both advanced and emerging market economies.

### IV. POLICY AND ANALYTICAL IMPLICATIONS (Implied by findings)
- Empirical implications:
  - A sizable global component exists, but global financial shocks explain about 20 to 40 percent of variation in domestic FCIs on average; domestic factors explain the remainder.
  - Domestic monetary policy retains relevance: domestic monetary policy shocks explain about 15 percent of variation for countries with flexible exchange rates.
  - Faster reaction of local financial conditions to global shocks compared with domestic policy rate changes implies challenges for timely policy responses.
- Policy considerations:
  - Monitoring and measuring FCIs purged of macroeconomic fundamentals can help distinguish non-fundamental spillovers from fundamentals-driven comovements.
  - Strengthening financial development may reduce sensitivity to global financial shocks.
  - Recognize that greater financial linkages increase exposure to global financial cycles; policymakers in financially integrated economies should take external factors into account when setting domestic policy.
  - Prudential tools may face leakage and limited efficacy in open economies; coordination and timely response frameworks may be required.

*Italic source attribution: wp1815 - References*

### 1. IMF Financial Stress Index  Versus Financial Conditions Index

### 1. IMF Financial Stress Index  Versus Financial Conditions Index

### Financial Conditions Indices: methodology and interpretation
- Uses a statistical dynamic factor model (DFM) following Gilbert and Meijer (2005) to generate unobservable (latent) factors from 43 country FCIs.
- The three-factor DFM representation: FCI_c,t = λ1,c x1,t + λ2,c x2,t + λ3,c x3,t (c and t denote country and time).
- Note: Higher values indicate tighter-than-average financial conditions.

### Three latent factors and their historical interpretation
- The three factors are characterized as:
  - “Emerging market” factor
  - “Euro area” factor
  - “Global financial crisis” factor
- Each factor spikes during the global financial crisis; the emerging market factor is notably tighter in the late 1990s and the euro area factor around 2012.

### Global factor, PCA, and relationship with U.S. measures
- A single global factor summarizes financial conditions across countries and is consistent with a global financial cycle.
- Robustness check: a single global factor computed via principal components analysis (PCA) moves in tandem with the DFM global factor.
- The single-factor global measures closely track the U.S. FCI and the VIX.
- Key statistic: The average correlation between the U.S. FCI and the two measures of global financial conditions and the VIX is 82 percent.
- Literature link: Findings align with Rey (2013) on U.S. predominance and roles of the U.S. dollar.

### Share of variance in country FCIs explained by global factors
- On average, global financial conditions account for about 30 percent of the variation in financial conditions across countries based on panel regressions of country FCIs on global measures.
- The share reaches almost 70 percent in several economies (not shown).
- The three-factor model explains greater than 40 percent of FCI variability on average.
- These magnitudes exceed Miranda-Agrippino and Rey (2015), who report about 21 percent for risky asset prices.

### Time variation and robustness of global influence
- Using 36- and 60-month rolling windows, the total variance explained by the first principal component across countries’ FCIs displays cyclical patterns but a broadly flat trajectory over the past 20 years.
- Patterns are robust to Forbes and Rigobon (2002)-type adjustments.
- Rolling-regression R2 statistics based on 36- and 60-month windows show broadly similar patterns.
- Regional robustness exercises (Figures 7 and 8) corroborate the lack of a sharp uptrend in the variance across countries’ FCIs explained by global financial conditions.
- Related empirical note: Bekaert and others (2016) document equity return correlations rising from the end of the 1990s through the global financial crisis, then declining notably.
- Citation from literature within the source: Carrieri, Chaieb, and Errunza (2013) argue that emerging markets are not yet effectively integrated with global markets.

### Regional and group-level variance-attribution findings (summary of figures)
- Figure 5: Displays variance accounted for by one- and three-factor models (In percent) across country groups; simple averages used.
- Figure 6: Shows variance attributable to global conditions, 1995–2016 (In percent), using 36- and 60-month rolling windows.
- Figure 7: Rolling three-year R-squared of domestic FCI on global FCI (one-factor model) by groups including All, Small open, advanced economies, Emerging market economies, Emerging Asia, Emerging Europe, Latin America.
- Figure 8: Rolling three-year R-squared using a three-factor model by regions (EM Asia, EM Latam, EM other, SOE). SOE denotes 39-small advanced and emerging market economies in the sample (excludes Germany, Japan, U.S., U.K.).

### Country characteristics and sensitivity to global financial conditions
- The U.S. FCI is used as a proxy for global financial conditions.
- Country characteristics examined as potential determinants of sensitivity include:
  - Financial linkages with the United States (foreign direct investment, banking, portfolio)
  - Financial openness and development
  - Institutional quality
  - Exchange rate regime
- References informing variable choice: Forbes and Chinn 2004; Aizenman, Chinn, and Ito 2015; Sahay and others 2015.

*Source: 1. IMF Financial Stress Index  Versus Financial Conditions Index*

### Appendix Table 2 lists the set of variables used and their sources.

### wp1815 - Appendix Table 2 lists the set of variables used and their sources.

### Overview and motivation
- Expectation: FCIs of countries that are more financially open and that feature stronger financial linkages with the United States should be more sensitive to global financial conditions. Conversely, countries with strong institutional and policy frameworks and deep financial markets should display less sensitivity.
- Empirical confirmation: Granger causality and convergent cross-mapping analyses indicate the U.S. FCI provides more statistically significant information about future FCIs in other countries than other financial centers, with an average p-value of 7 percent.
- Sample for panel regression: 39 advanced and emerging market economies from 1991 to 2016; countries that could be main drivers of global financial conditions (Germany, Japan, the United Kingdom, the United States) are excluded.

### Baseline panel regression specification
- Baseline equation (textual form preserved):
  - FCI_c,t = α_i + β1 FCI_t^US + β2 FCI_c,t−1 + β3 FCI_t^US * FCI_c,t−1 + β4 CCHAR_c,t−1 + ε_it
- Definition of variables:
  - FCI denotes domestic financial conditions.
  - Country characteristics (CCHAR) include measures of integration (trade and financial openness), linkages to the United States (FDI, banking, portfolio, trade), exchange rate flexibility, financial development, and rule of law.
  - Additional controls (Z): global variables (commodity price inflation and global growth) and domestic variables (growth, inflation, current account balance).
- Estimation features:
  - All explanatory variables except the global driver are lagged.
  - Model includes country fixed effects.
  - Standard errors clustered at the country level.

### Determinants of sensitivity to global financial shocks (summary of empirical findings)
- Direct effect of U.S. FCI:
  - Expected Sign: +
  - Estimated Sign: +
  - Significance: ***
- Interactions and signs (from Table 1 summary):
  - FDI linkages with the U.S.: Expected +, Estimated +, Significance ** 
  - Portfolio linkages with the U.S.: Expected +, Estimated −, Significance: not significant
  - Banking linkages with the U.S.: Expected +, Estimated −, Significance: not significant
  - Trade linkages with the U.S.: Expected +, Estimated +
  - Trade openness: Expected +, Estimated +, Significance ** 
  - Financial openness: Expected +, Estimated +
  - Exchange rate flexibility: Expected −, Estimated +
  - Financial development: Expected −, Estimated −, Significance ** 
  - Rule of law: Expected −, Estimated −
- Interpretation:
  - Financial linkages (especially FDI stock) with the U.S. are most closely associated with higher synchronization of domestic FCIs with global financial conditions.
  - Portfolio and bank linkages matter less robustly; signs are mixed across specifications.
  - Greater financial development and deeper financial market depth (equity, bond size and liquidity) attenuate the impact of global financial shocks.
  - Trade linkages to the U.S. do not matter materially; trade linkages with rest of world may matter (possibly capture indirect financial linkages).
  - No clear, robust role for exchange rate regime or general capital account openness; specific financial integration measures (FDI) are significant.

### Panel VAR setup for transmission analysis
- Baseline panel VAR variables (monthly, for 25 small open economies with flexible exchange rates, 2001–2016):
  - Y_c,t = [FCI_t^* , ∆IP_c,t , ∆CPI_c,t , FCI_c,t , Δi_c,t ]' where:
    - ∆IP = (log) industrial production growth,
    - ∆CPI = inflation based on log CPI,
    - Δi = change in domestic policy rate,
    - FCI_t^* = U.S. FCI (proxy for global financial conditions),
    - FCI_c,t = country-specific FCI.
  - X_t added for robustness: global industrial production growth, commodity prices, global interest rate measures (U.S. shadow rate proxies).
- Estimation details:
  - Four lags.
  - Pesaran, Shin, and Smith (1999) mean group estimator.
  - Impulse responses from Cholesky decompositions assuming U.S. FCI moves first and domestic interest rates move last.
  - Standard errors via nonparametric bootstrap with 1,000 replications.

### Impulse response findings (panel VAR)
- Global financial shocks:
  - A one standard deviation tightening in global financial conditions (U.S. FCI) raises domestic financial conditions by, on average, about one-fifth of a standard deviation across the small open economies with flexible exchange rates in the sample.
- Domestic monetary policy shocks:
  - Changes in local policy rates account for appreciable effects on local FCIs.
  - Local FCIs react faster and more strongly to global financial shocks than to changes in domestic policy rates — implying monetary policy may need to react very quickly and strongly to offset global shocks, with potential undesirable side effects.
- Robustness:
  - Results robust to alternative lag specifications, inclusion of global controls, using levels, and adding exchange rate terms.

### Peak impulse response robustness (Table 4 summary)
- Panel VAR impulse responses: Peak magnitudes (In percent, standard deviations) and Time to Peak (Months)
  - Exog: commodity price inflation
    - Monetary Policy Shock: Peak Magnitude 16.5; Time to Peak 8
    - Global Financial Shock: Peak Magnitude 15.6; Time to Peak 5
  - Exog: commodity price inflation, growth in global IP, change in U.S. shadow rate (Xia Wu)
    - Monetary Policy Shock: Peak Magnitude 14.8; Time to Peak 9
    - Global Financial Shock: Peak Magnitude 15.7; Time to Peak 6
  - Exog: commodity price inflation, growth in global IP, change in U.S. shadow rate (Krippner)
    - Monetary Policy Shock: Peak Magnitude 14.0; Time to Peak 9
    - Global Financial Shock: Peak Magnitude 15.2; Time to Peak 6
  - Endog: change in NEER
    - Monetary Policy Shock: Peak Magnitude 15.4; Time to Peak 9
    - Global Financial Shock: Peak Magnitude 20.1; Time to Peak 5
  - Endog: change in REER
    - Monetary Policy Shock: Peak Magnitude 15.5; Time to Peak 9
    - Global Financial Shock: Peak Magnitude 20.3; Time to Peak 5
  - Endog: change in bilateral exchange rate with the U.S.
    - Monetary Policy Shock: Peak Magnitude 15.4; Time to Peak 9
    - Global Financial Shock: Peak Magnitude 20.5; Time to Peak 5

### Individual-country VAR impulse response robustness (Table 5 summary)
- Peak impulse responses (In percent, standard deviations)
  - Small Open Advanced Economies:
    - Baseline specification: World FCI 15.3; Domestic Policy Rate 9.4; Domestic FCI 19.9
    - Short specification: World FCI 11.9; Domestic Policy Rate 9.7; Domestic FCI 17.7
    - Long specification: World FCI 12.9; Domestic Policy Rate 9.4; Domestic FCI 14.3
  - Emerging Markets:
    - Baseline specification: World FCI 16.3; Domestic Policy Rate 6.3; Domestic FCI 17.3
    - Short specification: World FCI 12.0; Domestic Policy Rate 6.1; Domestic FCI 14.4
    - Long specification: World FCI 13.9; Domestic Policy Rate 4.5; Domestic FCI 11.6
  - Lag sensitivity (Small Open Advanced Economies):
    - 1-quarter lag: World FCI 8.4; Domestic Policy Rate 7.0; Domestic FCI 14.4
    - 3-quarter lag: World FCI 12.0; Domestic Policy Rate 8.8; Domestic FCI 14.5
    - 4-quarter lag: World FCI 11.7; Domestic Policy Rate 9.3; Domestic FCI 13.8
    - 6-quarter lag: World FCI 11.1; Domestic Policy Rate 9.5; Domestic FCI 13.2
  - Lag sensitivity (Emerging Markets):
    - 1-quarter lag: World FCI 10.1; Domestic Policy Rate 4.1; Domestic FCI 11.4
    - 3-quarter lag: World FCI 12.1; Domestic Policy Rate 4.3; Domestic FCI 11.8
    - 4-quarter lag: World FCI 13.0; Domestic Policy Rate 4.6; Domestic FCI 11.7
    - 6-quarter lag: World FCI 13.6; Domestic Policy Rate 4.8; Domestic FCI 11.4

### Variance decomposition: shares attributable to global financial and monetary policy shocks
- Panel VAR (average for small open economies with flexible exchange rates):
  - About 21 percent of the variation in domestic FCIs is attributed to global financial shocks.
  - Domestic monetary policy shocks account for about 15 percent of the fluctuations in FCIs.
- Individual-country VAR averages:
  - Global financial conditions: about 40 percent of domestic FCI variations.
  - Monetary policy shocks: about 12 percent of domestic FCI variations.
- Statistical significance:
  - Variance decompositions statistically significant at the 10 percent level.
- Robustness (Table 6 summary; variance decomposition at a 12-month horizon; In percent):
  - Exog: commodity price inflation — Global Financial Conditions 15.6; Domestic Monetary Policy 15.5; Other 68.9
  - Exog: commodity price inflation, growth in global IP, change in U.S. shadow rate (Xia Wu) — Global 16.8; Domestic 11.8; Other 71.4
  - Exog: commodity price inflation, growth in global IP, change in U.S. shadow rate (Krippner) — Global 17.8; Domestic 12.5; Other 69.7
  - Endog: change in NEER — Global 13.9; Domestic 11.9; Other 74.2
  - Endog: change in REER — Global 15.1; Domestic 12.6; Other 72.3
  - Endog: change in bilateral exchange rate with the U.S. — Global 14.6; Domestic 12.2; Other 73.2
- Individual-country VAR variance decomposition summaries (Table 7; In percent):
  - Small Open Advanced Economies:
    - Baseline specification: World FCI 26.1; Domestic Policy Rate 20.1; Domestic FCI 28.4
    - Long specification: World FCI 23.0; Domestic Policy Rate 18.0; Domestic FCI 27.9
    - Short specification: World FCI 27.6; Domestic Policy Rate 26.4; Domestic FCI 46.0
  - Emerging Markets:
    - Baseline specification: World FCI 38.2; Domestic Policy Rate 11.3; Domestic FCI 25.6
    - Long specification: World FCI 36.4; Domestic Policy Rate 10.5; Domestic FCI 24.1
    - Short specification: World FCI 40.8; Domestic Policy Rate 18.0; Domestic FCI 41.3
  - Lag sensitivity (Small Open Advanced Economies, In percent):
    - 1-quarter lag: World FCI 21.8; Domestic Policy Rate 23.4; Domestic FCI 40.8
    - 3-quarter lag: World FCI 26.2; Domestic Policy Rate 22.2; Domestic FCI 35.6
    - 4-quarter lag: World FCI 25.8; Domestic Policy Rate 23.4; Domestic FCI 32.5
    - 6-quarter lag: World FCI 25.0; Domestic Policy Rate 21.1; Domestic FCI 31.1
  - Lag sensitivity (Emerging Markets, In percent):
    - 1-quarter lag: World FCI 31.7; Domestic Policy Rate 14.6; Domestic FCI 35.2
    - 3-quarter lag: World FCI 39.1; Domestic Policy Rate 13.6; Domestic FCI 32.9
    - 4-quarter lag: World FCI 41.2; Domestic Policy Rate 14.1; Domestic FCI 29.5
    - 6-quarter lag: World FCI 38.9; Domestic Policy Rate 13.4; Domestic FCI 27.5

### Country heterogeneity
- Country-by-country VARs show considerable variation in the share of domestic FCI fluctuations attributable to global financial shocks.
- Distributional patterns:
  - Histogram of shares is broadly bi-modal when including a broader set of countries.
  - Small open advanced economies cluster in the left tail (lower shares attributable to global financial shocks).
  - Emerging market economies generally exhibit a greater proportion of FCI variability explained by global financial conditions; in some cases the proportion exceeds 60 percent.
- Financial integration correlation:
  - Countries relatively more financially integrated with the rest of the world exhibit higher shares of FCI variability due to global financial shocks; differences larger for emerging markets.

### More refined identification of monetary policy shocks
- Motivation: Precisely identifying monetary policy shocks in VARs is difficult; complementary identification improves inference.
- Methodology (three-step, Gertler and Karadi (2015) approach):
  1. Identify monetary policy shocks as movements in various interest rates within a narrow high-frequency window (e.g., 30 minutes) following central bank policy announcements; construct monthly monetary policy surprise series by applying a scaling factor to high-frequency shocks.
  2. Use constructed monetary policy surprises as instruments/regressors for the monetary policy indicator (one-year or two-year government bond yields depending on data availability); fitted values correspond to policy indicator changes attributable to monetary policy shocks.
  3. Use fitted values in original VAR to derive the column in the coefficient matrix corresponding to the monetary policy shock and compute impulse responses.
- Country cases with identified shocks: Australia, New Zealand, Norway, Sweden (chosen for floating regimes, deep financial markets, and data availability).
- Result: VARs using these identified monetary policy shocks yield impulse responses of domestic FCIs to monetary policy shocks that are very similar and statistically indistinguishable from responses using Cholesky decomposition for these four countries.
- Note: For these four country case studies, share of FCI variation characterized by fluctuations in global financial conditions and domestic monetary policy is, on average, 15 percent and 33 percent, respectively (based on a Cholesky decomposition).

### Temporal comparison: 2001–07 versus 2010–16
- Finding: No discernible change in the importance of global financial conditions in influencing local FCIs over time.
- Variance decompositions for 2001–07 and 2010–16 are broadly stable and not statistically different at the 95 percent level.

*Source: wp1815 - Appendix Table 2 lists the set of variables used and their sources (authors’ estimates and tables described in the source document).*

### 1. Australia

### 1. Australia

### Summary of empirical findings
- A single factor—global financial conditions—summarizes the dynamics of a significant share of financial conditions around the world and moves in tandem with the FCI of the U.S. and standard measures of global risk such as the VIX.
- The fraction of fluctuations in countries’ domestic financial conditions attributed to global financial conditions does not appear to have increased markedly over the past two decades.
- Stronger financial linkages with the U.S. increase the sensitivity of domestic financial conditions to global financial shocks.
- Financial development attenuates the sensitivity of domestic financial conditions to global financial shocks.
- On average, countries are still able to steer their domestic financial conditions to a significant extent despite the influence of global financial conditions.
- Domestic financial conditions respond faster and more strongly to global financial shocks than to changes in the domestic monetary policy stance, making timely and effective policy reactions often challenging.

### Evidence and measurement notes
- The panel vector autoregression model is used to decompose the share of domestic FCI fluctuations attributable to global financial shocks for precrisis (2001–07) and postcrisis (2010–16) samples; the figure reports squares for point estimates and lines for the 90 percent confidence bands.
- Figure 13 (Share of Domestic Financial Index Fluctuations Attributable to Global Financial Shocks Before and After the Global Financial Crisis) displays results for countries in the sample with flexible exchange rates, including Australia.

### Policy implications
- Policymakers should account for the strong and rapid transmission of global financial shocks to domestic financial conditions when designing response strategies.
- Given the faster domestic response to global shocks relative to domestic policy actions, authorities may need to prioritize speed and preemptive elements in policy design to mitigate spillovers.
- Financial development can serve as a buffer against global shock transmission and may be a policy objective to enhance resilience.

### Suggested avenues for future research (policy-relevant)
- Investigate asymmetries and threshold effects: do countries’ financial conditions react differently to a tightening compared to a loosening of global financial conditions?
- Examine whether temporary, acute episodes of global tightening are more influential for countries’ FCIs than protracted periods of accommodative global financial conditions.
- Further explore transmission channels across countries, including the role played by different types of financial intermediaries.

*Source: Authors’ estimates and analysis (panel VAR; precrisis 2001–07 and postcrisis 2010–16; 90 percent confidence bands).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp1815.pdf_