## _wp14109

## Source details

**Canonical URL:** [_wp14109](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14109.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14109.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14109.pdf.json)

---

### I. Introduction — context and headline market responses
- Ultra loose monetary policy in advanced economies prompted a global search for yield into emerging markets (EMs) after the Global Financial Crisis; May 2013 Fed "tapering" talk triggered market volatility across EMs.
- Headline impacts on five especially affected countries (Brazil, India, Indonesia, Turkey, South Africa) during May 22–(end of) August 2013:
  - bond yields rise by 2½ percentage points
  - equity market fall by 13¾ percent
  - exchange rates depreciate by 13½ percent
  - reserves declined by 4.1 percent
- Aggregate summary for May 22–end-June 2013 (21 countries in sample):
  - currencies across EMs depreciated 3 percent (on average)
  - spreads rose 1 percentage points
  - equities fell 7 percent
- Empirical focus: systematic analysis of market reaction in 21 EMs from January 1, 2013, to January 22, 2014 following FOMC announcements; examines role of macro fundamentals, financial depth and integration, trade linkages with China, capital-flow measures, and macroprudential policies.

### II. Market phases and expectations
- Market expectation revisions (New York Fed surveys):
  - April survey: most conservative expected Fed to buy US$85 billion a month through July and reduce to US$55 billion at December meeting.
  - July survey (after Bernanke speech): markets expected US$65 billion per month at September meeting and reduce to US$45 billion after December meeting.
- Market phases:
  - Phase 1 (May 21 to end-Jun 2013): broad-based sell-off across EMs (aggregate stats above).
  - Phase 2 (Jul to end-Dec 2013): greater differentiation based on fundamentals; investors focused on India, Indonesia, Brazil, Turkey, South Africa.
  - Phase 3 (Jan to Mar 21, 2014): another bout of volatility triggered by China shadow banking news and idiosyncratic country events; clearer Fed communication reduced market impact of actual tapering.
- Exposure to China:
  - Exposure to China can provide buffers when volatility stems from FOMC announcements; but when negative news emanates from China, exposure to China does not serve as a significant buffer.

### III. Data, sample, and variables
- Sample: 21 emerging markets — Brazil, Chile, China, Colombia, Hungary, India, Indonesia, Israel, Korea, Malaysia, Mexico, Peru, Philippines, Poland, Romania, Russia, Singapore, South Africa, Taiwan, Thailand, and Turkey.
- Time coverage: daily data from January 1, 2013 to January 22, 2014.
- Dependent variables:
  - exchange rates (local currency/US$)
  - local currency government bond yields (10-year and five-year)
  - equity prices
- Country characteristic categories:
  - (i) macro fundamentals: inflation, fiscal balance/GDP, current account/GDP, reserves/GDP
  - (ii) financial depth: bank credit/GDP, M2/GDP, M3/GDP, stock market capitalization/GDP, bid-ask spread in 10-year local currency bond market
  - (iii) EM growth prospects: current and one-year ahead growth forecasts
  - (iv) China exposures: China one-year ahead growth and trade linkage with China
  - (v) financial integration: portfolio assets and liabilities/GDP (last quarter 2012)
  - (vi) excess capital inflows relative to benchmark (cumulated up to last quarter 2012)
  - (vii) capital flow measures and macroprudential policies (cumulated from 2000 to last quarter of 2012; indices take values -1, 0, +1)
- Lagging: categories (i)-(v) lagged one-quarter; (v)–(vii) treated as time-invariant over sample window.

### IV. Empirical methodology
- Event window: two-day window around FOMC meetings and release of minutes (one day before and one day after); robustness check with four-day window yields similar results.
- Event classification: negative, positive, or non-events. Negative events defined by exchange rate depreciation, bond yield increases, equity price declines; rule of thumb: at least two coefficients need to be statistically significant to classify an event as negative or positive.
- Events analyzed: 17 Fed-policy-related events during January 1, 2013 to January 22, 2014.
- Two-step strategy:
  - Step 1: pooled regressions of two-day changes on event dummies (Equation (1)).
  - Step 2: pooled events across 21 countries with interactions between negative-event dummies and country characteristics (Equation (2)); country fixed effects included.
- Identification advantages: narrow two-day window focuses on EM-specific reactions and explicit interactions to capture differentiation.

### V. Classification of events and pooled reaction magnitudes
- Markets reacted negatively to meetings (4) and (8), and minutes (3), (5), and (7).
- Markets did not react negatively to the December meeting (meeting (9)) when tapering was announced (tapering likely already priced in).
- Positive events included meetings (6) and (7), and minutes (4) and (6); strong positive reaction after September 18 meeting (meeting (6)) when Fed postponed tapering.
- Pooled regression summary statistics (two-day changes, Table 2):
  - Observations: 357 (for each dependent variable)
  - R-squared: Exchange rate 0.615; Government bond yields 0.358; Stock prices 0.418
  - Example coefficients (two-day changes, percent):
    - meeting_4: Exchange rate 1.94*** (0.229); Government bond yields 0.24*** (0.046); Stock prices -3.02*** (0.311)
    - meeting_6: Exchange rate -1.60*** (0.165); Government bond yields -0.20*** (0.043); Stock prices 1.50*** (0.381)
    - minutes_3: Exchange rate 0.47*** (0.099); Government bond yields 0.09*** (0.019); Stock prices -0.78*** (0.236)
  - Standard errors clustered at country-level; ***, **, * denote statistical significance at 1, 5, 10 percent respectively.

### VI. Determinants of FX reactions — macro fundamentals
- Better fundamentals associated with smaller FX depreciations around negative events:
  - Higher current account balances, lower inflation, higher reserves/GDP, and better fiscal positions reduced depreciation magnitudes.
  - Quantitative example: a country with a one percent of GDP higher-than-average current account balance would see its exchange rate depreciate by 0.03 percentage points less than the average over a two-day period.
- Table 3 interaction coefficients (exchange rate reactions, two-day window):
  - Interactions with CA/GDP: -0.031** (0.011)
  - Reserves/GDP: -0.011*** (0.003)
  - Inflation: 0.061** (0.029)
  - Fiscal balance/GDP: -0.053** (0.020)
  - Non-interacted terms largely not statistically significant around non-negative events.

### VII. Financial depth, market structure, and resilience
- Deeper domestic financial markets mitigated exchange rate depreciation:
  - Table 4 interaction coefficients (2-day changes in exchange rate):
    - Interactions with Bank credit/GDP: -0.005** (0.002)
    - M3/GDP: -0.004*** (0.001)
    - M2/GDP: -0.005*** (0.001)
    - Stock market cap/GDP: -0.004** (0.001)
    - Bid-ask spread: -0.253 (1.430)
  - Non-interacted financial-depth coefficients generally negative and often significant (e.g., Non-interacted M2/GDP: -0.095*** (0.026)).
- Economic significance: effects can reach up to 0.3 percentage points depreciation over two days for one standard deviation increases in vulnerability.

### VIII. Growth prospects and exposure to China
- Growth prospects:
  - Better current or one-year-ahead growth forecasts associated with smaller FX depreciations.
  - Table 5 interactions:
    - Real GDP growth interaction: -0.082* (0.045)
    - Annual growth forecast - 1 yr ahead: -0.118* (0.065)
- Exposure to China:
  - Interaction between negative event dummy and exposure to China (exports+imports to China/GDP) is negative and statistically significant in several specifications (Table 6).
  - Chart regression line example: y = -0.5115x + 22.968, R² = 0.3041.
  - China growth alone did not have a consistent significant impact; triple interactions with low China growth were not statistically significant.

### IX. Capital flow measures, financial integration, and macroprudential policies
- Capital flow measures (CFMs) and integration:
  - Countries that imposed CFMs amid large inflows and those with greater international financial integration experienced smaller FX depreciations (Table 7).
  - Triple interaction (dummy × international financial integration × CFM) negative and statistically significant.
  - Selected interaction coefficients:
    - Excess capital flows * CFM: -0.001** (0.000)
    - International financial integration: -0.002* (0.001)
    - International financial integration * CFM: -0.001* (0.000)
- Macroprudential policies and depth:
  - Countries that tightened macroprudential policies prior to 2013 experienced less FX depreciation (Table 8).
  - But macroprudential benefits diminish with greater financial depth:
    - Macro prudential: -0.023** (0.009); -0.052*** (0.006) in different specs
    - Stock market cap/GDP: -0.008*** (0.001)
    - Macro prudential * Stock market cap: 0.001*** (0.000) (positive interaction)
  - Interpretation: CFMs more effective with greater integration; macroprudential measures effective but marginal benefits decline as financial depth increases.

### X. Bond yields and equity reactions
- Bond yields:
  - Regression results for two-day changes in 10-year bond yields broadly mirror FX findings: stronger fundamentals, deeper markets, larger exposure to China, and tighter macroprudential stances associated with smaller yield increases.
  - Example magnitude: a one standard deviation higher reserves/GDP or lower inflation is associated with 4 basis points larger increase in yields over a two-day period (reported as association).
- Equity prices:
  - Little significant association between equity changes and most country characteristics (Table 10); equity investors differentiated less across countries.
  - Exception: interaction of negative event dummy with excess capital inflows is negative and statistically significant—countries with larger inflows prior to the episodes experienced larger equity declines.

### XI. Joint specification and salient characteristics
- Joint interactions (Table 11) yield weaker but present evidence of differentiation.
- Macro fundamentals that matter in multivariate specs include:
  - Reserves/GDP
  - Growth
  - Inflation
  - Change in current account
  - Foreign-exchange-denominated debt (stock)
- Selected coefficients (Table 11, preserved exactly where reported):
  - dummy coefficients: 0.946*, 2.905***, 0.327**, 0.489**, -1.414, -2.342 (columns 1–6 respectively).
  - Reserves/GDP interaction: -0.004; -0.027***; -0.003*; -0.004; 0.024; 0.038 (selected entries).
  - Macro prudential interaction: 0.011; -0.031**; -0.009***; -0.016***; -0.012; 0.057** (selected entries).

### XII. Principal component — overall vulnerability/resilience
- Constructed a resilience index by normalizing and standardizing country characteristics (higher values = lower vulnerability) and taking the first principal component.
- Most resilient EMs: Singapore, Taiwan, Malaysia, Korea, Thailand.
- Table 12 (2-day changes):
  - Exchange rate: Dummy 1.032***; Dummy*Resilience -0.087***; Resilience 0.102; Observations 306; R-squared 0.249.
  - Government bond yields: Dummy 0.130***; Dummy*Resilience -0.017*; Resilience 0.054; Observations 306; R-squared 0.187.
  - Stock prices: Dummy -1.589***; Dummy*Resilience -0.008; Resilience 0.070; Observations 306; R-squared 0.212.
- Interpretation: higher resilience reduces FX depreciation and yield increases around negative events; resilience less relevant for stock price reactions.

### XIII. Robustness checks and alternative event measures
- Alternative events: changes in expected future short-term rates (euro-dollar futures, OIS, Fed Fund futures) yield broadly similar differentiation results.
- Table 13 (2-day change in euro-dollar futures (4 years ahead)) reports many significant coefficients across 17 columns (examples preserved exactly in source).
- Revision in market expectations of Fed purchases (RΔ variable) based on New York Fed surveys:
  - June survey revision: markets revised expected Fed purchases downward by US$20 billion for September relative to April expectations.
- Table 14 (FX reaction to revision of market expectations of FED purchases, 2-day changes) reports negative and significant main coefficients across columns and interactions:
  - Revision in market expectation: -0.00908***; -0.01149***; -0.00802**; -0.00410**; -0.00960***; -0.01370***; -0.01373***; -0.01417***; -0.01257***; -0.00948***; -0.01135***; -0.00850**; -0.00838; -0.01065***; -0.00816***; -0.01079***; -0.01272*** (columns 1–17).
  - Interaction examples:
    - Inflation interaction: -0.00140** [0.001]
    - Bank credit/GDP interaction: 0.00006** [0.000]
    - M3/GDP interaction: 0.00004* [0.000]
    - M2/GDP interaction: 0.00006*** [0.000]
    - Macro prudential interaction: -0.00053** [0.000]
    - Capital flow measures interaction: 0.00032* [0.000]
- Robustness: classification of negative events and main inferences robust to longer event windows and alternative sample spans (e.g., all FOMC announcements 2009–2013).

### XIV. Key quantitative summary (selected statistics preserved exactly)
- Table 2 pooled regressions: Observations 357; R-squared: Exchange rate 0.615; Government bond yields 0.358; Stock prices 0.418.
- Table 3 exchange-rate dummy coefficients: Dummy: 1.030***, 1.373***, 0.800***, 0.956*** (standard errors: (0.091), (0.133), (0.170), (0.084)); Observations: 357; R-squared: 0.234, 0.241, 0.233, 0.236.
- Table 4 financial depth interactions: Dummy: 1.442***, 1.585***, 1.450***, 1.335***, 0.981*** (standard errors: (0.178), (0.164), (0.123), (0.173), (0.132)); Interactions with Bank credit/GDP: -0.005** (0.002); M3/GDP: -0.004*** (0.001); M2/GDP: -0.005*** (0.001); Stock market cap/GDP: -0.004** (0.001); Observations vary by column (e.g., 357, 289, 357, 357, 340).
- Appendix A1 summary statistics (Obs, Mean, Std. Dev., Min, Max):
  - Change in FX (1-day before and after): Obs: 357; Mean: 0.16; Std. Dev.: 0.99; Min: -3.10; Max: 3.65.
  - Change in Yields (1-day before and after): Obs: 357; Mean: 0.02; Std. Dev.: 0.15; Min: -0.80; Max: 0.84.
  - Change in Equity (1-day before and after): Obs: 357; Mean: -0.11; Std. Dev.: 1.63; Min: -5.96; Max: 6.26.
  - Reserves/GDP: Obs: 357; Mean: 30.78; Std. Dev.: 21.48; Min: 10.12; Max: 93.71.
  - M2/GDP: Obs: 357; Mean: 79.01; Std. Dev.: 57.83; Min: 18.48; Max: 243.49.
  - Stock Market Capitalization/GDP: Obs: 357; Mean: 75.31; Std. Dev.: 53.27; Min: 8.16; Max: 223.70.
  - Cummulative Macroprudential Index: Obs: 340; Mean: 6.90; Std. Dev.: 8.67; Min: -1.00; Max: 40.00.
  - Excessive Capital Inflow: Obs: 272; Mean: 24.59; Std. Dev.: 39.34; Min: -52.36; Max: 107.25.
  - Export and Import to China/GDP: Obs: 340; Mean: 9.80; Std. Dev.: 8.47; Min: 1.60; Max: 33.21.

### XV. Conclusions and policy-relevant findings
- The Fed’s plans and decision to taper were accompanied by significant market reaction in many EMs.
- Markets differentiated across countries during episodes of turmoil based on country characteristics, including macroeconomic fundamentals and economic and financial structures.
- Countries with better fundamentals and greater financial depth were less hit than others.
- Countries that had tighter macroprudential policies and had resorted to capital flow measures prior to May 2013 fared better.
- Strong trade ties with China can help dampen market reaction when no bad news emanates from China; when bad news from China coincides with the initial reason for turmoil in EMs, strong ties with China do not dampen market reactions.

*Source: IMF Staff working paper content provided in the uploaded PDF _wp14109.*

### 1. List of FOMC Meetings and Minutes ........................................................................... 9

### _wp14109 - 1. List of FOMC Meetings and Minutes ........................................................................... 9

### Major sections (table of contents)
- 1. List of FOMC Meetings and Minutes ........................................................................... 9
- 2. Market Reactions to FOMC Meeting and Release of Minutes ................................... 13
- 3. FX Reaction and Macroeconomic Fundamentals ....................................................... 15
- 4. FX Reaction and Financial Depth ............................................................................... 19
- 5. FX Reaction and Emerging Market Growth Forecasts ............................................... 18
- 6. FX Reaction and Emerging Market, China's Growth and Exposure to China ........... 19
- 7. FX Reaction, Financial Integration, Capital Flows and Capital Flow Measures........ 20
- 8. FX Reaction, Financial Depth and Macro Prudential Measures ................................. 20
- 10. Changes in Stock Prices and Macroeconomic Fundamentals ..................................... 24
- 11. Which Country Characteristics Matter the Most ........................................................ 26
- 12. Market Reaction and Country Resilience Principal Component ................................ 27
- 13. Reaction to Tightening and Country Characteristics .................................................. 30
- 14. Reaction to Revision of Market Expectations of FED Purchases and Country  
  Characteristics ....................................................................................................... 27

### Organizational focus implied by section headings
- FOMC meeting schedule and minutes (section 1)
- Market and FX reactions to FOMC communications and minutes (sections 2–8, 14)
- Links between FX/stock market responses and country-level fundamentals, financial depth, integration, capital flows, macroprudential measures, and exposure to China (sections 3–8, 10)
- Cross-country variation and determinants of market responses, including principal component analysis of country resilience and specific characteristic impacts (sections 11–13)

*Source: _wp14109 - 1. List of FOMC Meetings and Minutes ........................................................................... 9*

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

### _wp14109 - References .............................................................................................................

### I. Introduction — context and headline market responses
- Ultra loose monetary policy in advanced economies prompted a global search for yield into emerging markets (EMs) after the Global Financial Crisis; May 2013 Fed "tapering" talk triggered market volatility across EMs.
- Headline impacts on five especially affected countries (Brazil, India, Indonesia, Turkey, South Africa) during May 22–(end of) August 2013:
  - bond yields rise by 2½  percentage points
  - equity market fall by 13¾ percent
  - exchange rates depreciate by 13½ percent
  - reserves declined by 4.1 percent
- Since end-2013, additional bursts of market pressure occurred related in part to the Fed’s decision to taper and other factors (e.g., China’s shadow banking risks, idiosyncratic developments in Argentina, rising political and policy risks in Turkey).
- Aggregate summary for May 22–end-June 2013 (21 countries in sample):
  - currencies across EMs depreciated 3 percent (on average)
  - spreads rose 1 percentage points
  - equities fell 7 percent
- Empirical focus: systematic analysis of market reaction in 21 EMs from January 1, 2013, to January 22, 2014 following FOMC announcements; examines role of macro fundamentals, financial depth and integration, trade linkages with China, capital-flow measures, and macroprudential policies.

### II. What happened in the summer of 2013? — market expectations and phases
- New York Fed surveys of markets’ expectations of Fed bond purchases showed downward revisions after Chairman Bernanke’s May 22, 2013 speech; examples from surveys:
  - April survey: most conservative expected Fed to buy US$85 billion a month through July and reduce to US$55 billion at December meeting.
  - July survey (after Bernanke speech): markets expected US$65 billion per month at September meeting and reduce to US$45 billion after December meeting.
- Market phases and differentiation:
  - Phase 1 (May 21 to end-Jun 2013): broad-based sell-off across EMs (see aggregate stats above).
  - Phase 2 (Jul to end-Dec 2013): greater differentiation based on fundamentals (external and macro imbalances); investors focused on India, Indonesia, Brazil, Turkey, South Africa.
  - Phase 3 (Jan to Mar 21, 2014): another bout of volatility triggered by a confluence of factors including China’s shadow banking news and idiosyncratic country events; clearer Fed communication reduced market impact of actual tapering.
- Exposure to China:
  - Exposure to China can provide buffers when volatility stems from FOMC announcements; but when negative news emanates from China, exposure to China does not serve as a significant buffer.

### III. Data — sample, variables, and measurement
- Sample: 21 emerging markets — Brazil, Chile, China, Colombia, Hungary, India, Indonesia, Israel, Korea, Malaysia, Mexico, Peru, Philippines, Poland, Romania, Russia, Singapore, South Africa, Taiwan, Thailand, and Turkey.
- Time coverage: daily data from January 1, 2013 to January 22, 2014.
- Dependent variables:
  - exchange rates (local currency/US$)
  - local currency government bond yields (10-year and five-year)
  - equity prices
- Country characteristics (categories and examples):
  - (i) macro fundamentals: inflation, fiscal balance/GDP, current account/GDP, reserves/GDP
  - (ii) financial depth: bank credit/GDP, M2/GDP, M3/GDP, stock market capitalization/GDP, bid-ask spread in 10-year local currency bond market
  - (iii) EM growth prospects: current and one-year ahead growth forecasts
  - (iv) China exposures: China one-year ahead growth and trade linkage with China
  - (v) financial integration: stock of portfolio assets and liabilities/GDP (measured in last quarter of 2012)
  - (vi) excess capital inflows relative to a pre-specified benchmark (cumulated up to last quarter of 2012)
  - (vii) capital flow measures and macroprudential policies (cumulated from 2000 to last quarter of 2012; indices take values -1, 0, +1 for loosened/unchanged/tightened)
- Lagging and timing:
  - For variables in categories (i)-(v) data are lagged one-quarter.
  - Financial integration and cumulative indices (v)–(vii) are time-invariant over the sample window as used.

### IV. Empirical methodology — event study and identification
- Event window and identification:
  - Two-day window around FOMC meetings and release of minutes (one day before and one day after the event). Robustness: four-day window (Monday–Friday) yields qualitatively similar results.
  - Events are classified as negative, positive, or non-events. Negative events: exchange rate depreciation, bond yield increases, equity price declines. Rule of thumb: at least two of the coefficients need to be statistically significant to classify an event as negative or positive.
- Events analyzed:
  - 17 news events related to Fed monetary policy during January 1, 2013 to January 22, 2014 (listed in Table 1).
- Two-step empirical strategy:
  - Step 1: pool events and document market reactions using regressions of two-day changes on event dummies (Equation (1)).
  - Step 2: pool the 17 events across 21 countries and estimate interactions between negative-event dummies and country characteristics (Equation (2)); include country fixed effects.
- Methodological features distinguishing this paper:
  - Narrow two-day window focused on EMs (avoids confounding events).
  - Identification of negative events explicitly.
  - Interactions of negative-event dummies with a wide set of country characteristics to gauge differentiation.
- Notes on measurement limitations:
  - Bond market depth proxies used (bank credit, M2, M3, stock market capitalization, bid-ask spread) because detailed sovereign bond market development measures are not available across the wide sample.
  - Foreign exchange interventions used as robustness control (sample reduced due to limited intervention data), with qualitatively similar results when included.

### V. Results — classification of events and determinants of market reactions
- Event classification and aggregate regression results (pooled, Equation (1), Table 2):
  - Markets reacted negatively to meetings (4) and (8), and minutes (3), (5), and (7).
  - Markets did not react negatively to the December meeting (meeting (9)) when tapering was announced—possible reason: tapering had already been priced in.
  - Positive events included meetings (6) and (7), and minutes (4) and (6); strong positive reaction after September 18 meeting (meeting (6)) when Fed postponed tapering.
  - Table 2 summary statistics (pooled regressions of 2-day changes):
    - Observations: 357 (for each dependent variable)
    - R-squared: Exchange rate 0.615; Government bond yields 0.358; Stock prices 0.418
    - Example significant coefficients from Table 2 (two-day changes, percent):
      - meeting_4: Exchange rate 1.94*** (0.229); Government bond yields 0.24*** (0.046); Stock prices -3.02*** (0.311)
      - meeting_6: Exchange rate -1.60*** (0.165); Government bond yields -0.20*** (0.043); Stock prices 1.50*** (0.381)
      - minutes_3: Exchange rate 0.47*** (0.099); Government bond yields 0.09*** (0.019); Stock prices -0.78*** (0.236)
    - Standard errors clustered at the country-level (values in parentheses). ***, **, and * denote statistical significance at the 1, 5, and 10 percent levels respectively.
- Determinants of FX market differentiation (summary of main findings):
  - Macroeconomic fundamentals:
    - Countries with higher current account balances, lower inflation, higher reserves (relative to GDP), and better fiscal positions saw smaller exchange rate depreciations and fared better around periods of turmoil.
    - Quantitative example: a country with a one percent of GDP higher-than-average current account balance would see its exchange rate depreciate by 0.03 percentage points less than the average over a two-day period.
  - Financial depth and market structure:
    - Countries with deeper domestic financial markets (larger size/liquidity) were less affected—liquidity allowed investors to move large amounts of capital without large price changes.
    - However, theoretical ambiguity exists because deeper systems can also allow faster unwinding of positions that could amplify instability.
  - Financial integration:
    - Ambiguous a priori: higher financial integration (portfolio assets+liabilities/GDP measured in last quarter 2012) could increase exposure to external shocks or provide diversification benefits; empirical analysis separates depth and integration effects.
  - Trade linkages with China:
    - Strong trade linkages with China can buffer EMs when volatility stems from FOMC announcements; but if China-originating negative news occurs, linkages can amplify vulnerability.
  - Capital flow measures and macroprudential policies:
    - Countries that tightened capital flow measures and maintained a tighter macroprudential stance prior to tapering talks exhibited more muted market reactions—suggesting policy stance affected the composition and volatility of flows.
  - Excess capital inflows:
    - Positive relationship observed between excess capital inflows (overflow) into EMs and bond market reactions (May 22–September 17, 2013): countries with larger excess inflows in the run-up to 2013 experienced larger bond yield increases during the May–Sep 2013 volatility episode.
- Equity markets:
  - The paper finds little significant association of equity price changes around FOMC announcements with country characteristics.
- Robustness and extensions:
  - Classification of negative events is robust to using a longer time span (all FOMC announcements 2009–2013).
  - Results are robust to using a four-day event window.

*Italic: Source — IMF Staff working paper content provided in the uploaded PDF _wp14109 - References.*

### 0.39 percent of GDP,  the estimated coefficient would imply the Turkish lira would

### _wp14109 - 0.39 percent of GDP,  the estimated coefficient would imply the Turkish lira would

### Exchange-rate reaction magnitudes around negative events
- The estimated coefficient implies the Turkish lira would depreciate an additional by 0.2 percentage points over the two-day window (around 30 percent annualized).
- Table 3 (Dependent Variable: 2-day Changes in Exchange Rate (Local Currency/US$) Pre and Post Event (in percent)) reports dummy coefficients:
  - Dummy: 1.030***, 1.373***, 0.800***, 0.956*** (standard errors: (0.091), (0.133), (0.170), (0.084))
- Observations in Table 3: 357, R-squared: 0.234, 0.241, 0.233, 0.236 (by column)
- Notes: Events defined as Meetings: June 19th, October 30th, Release of Minutes: May 22nd, August 21st, November 20th. All regressions include country fixed effects. Standard errors clustered at country-level. ***, **, and * denote statistical significance at 1, 5, and 10 percent respectively.

### Role of macroeconomic fundamentals
- Coefficient on non-interacted terms (macroeconomic fundamentals) around non-negative events are not statistically significant — i.e., little differentiation on the basis of fundamentals around announcements not characterized by turmoil.
- Interaction with positive events: estimated coefficient is of opposite sign and statistically significant, indicating that better fundamentals are associated with dampened market reactions in either direction.
- Specific interaction coefficients reported in Table 3:
  - Interactions with CA/GDP: -0.031** (0.011)
  - Reserves/GDP: -0.011*** (0.003)
  - Inflation: 0.061** (0.029)
  - Fiscal balance/GDP: -0.053** (0.020)
  - Non-interacted CA/GDP: 0.043 (0.068)
  - Non-interacted Reserves/GDP: 0.050 (0.044)
  - Non-interacted Inflation: 0.024 (0.056)
  - Non-interacted Fiscal balance/GDP: 0.007 (0.013)

### Financial depth and market resilience
- Financial depth tends to enhance resilience: countries with deeper financial markets experienced smaller exchange rate depreciations (Table 4).
- Mexico: example of deeper financial markets facilitating adjustment and showing little volatility after summer 2013 episodes.
- Table 4 interaction coefficients (Dependent Variable: 2-day Changes in Exchange Rate):
  - Dummy: 1.442***, 1.585***, 1.450***, 1.335***, 0.981*** (standard errors: (0.178), (0.164), (0.123), (0.173), (0.132))
  - Interactions with Bank credit/GDP: -0.005** (0.002)
  - M3/GDP: -0.004*** (0.001)
  - M2/GDP: -0.005*** (0.001)
  - Stock market cap/GDP: -0.004** (0.001)
  - Bid-ask spread: -0.253 (1.430)
  - Non-interacted Bank credit/GDP: -0.074*** (0.025)
  - Non-interacted M3/GDP: -0.090*** (0.023)
  - Non-interacted M2/GDP: -0.095*** (0.026)
  - Non-interacted Stock market cap/GDP: 0.015* (0.007)
  - Non-interacted Bid-ask spread: -3.518*** (1.096)
- Observations and R-squared (by column): Observations: 357, 289, 357, 357, 340. R-squared: 0.250, 0.286, 0.264, 0.242, 0.231.

### Growth prospects and exposure to China
- Countries with better growth prospects (actual level of activity in the quarter preceding the event or one-year-ahead growth forecasts) on average fared better; downward revisions in growth prospects associated with larger exchange-rate depreciations.
- Example: growth prospects for Brazil, Turkey, South Africa, and India were revised downward by about ¼ percentage points in 2013-14.
- Exposure to China (exports to + imports from China as ratio of GDP) acted as a stabilizing buffer in 2013:
  - Interaction between negative event dummy and exposure to China is negative and statistically significant (Table 6).
  - Chart regression line: y = -0.5115x + 22.968, R² = 0.3041 (figure relating "Score of 'most hit'" and Growth Revisions / Trade link to China).
- Table 5 interaction coefficients (Dependent Variable: 2-day Changes in Exchange Rate):
  - Dummy: 1.321***, 1.515***, 1.054*** (standard errors: (0.185), (0.285), (0.231))
  - Interactions with Real GDP growth: -0.082* (0.045)
  - Annual growth forecast - 1 yr ahead - Consensus: -0.118* (0.065)
  - Annual growth forecast - current -- Consensus: -0.009 (0.056)
  - Non-interacted Real GDP growth: 0.008 (0.019)
  - Non-interacted Annual growth forecast - 1 yr ahead: 0.416*** (0.090)
  - Non-interacted Annual growth forecast - current: 0.103 (0.074)
  - Observations: 357, R-squared: 0.233, 0.249, 0.229.
- China-specific interactions (Table 6):
  - Dummy: 1.060, 1.348***, 1.764, 2.715 (standard errors: (1.933), (0.116), (2.085), (3.169))
  - Exposure to China: -0.028***, -0.028***, -0.126 (standard errors: (0.007), (0.007), (0.174))
  - China's Annual growth forecast - 1 yr ahead - Consensus: 0.757***, 0.890***, 0.951*** (standard errors: (0.114), (0.141), (0.183))
  - China Growth * Exposure to China: 0.013 (0.022) and -0.007 (0.016) reported in different specifications
  - Observations: 357, 340, 340, 340. R-squared: 0.273, 0.251, 0.301, 0.302.
- China’s growth alone did not have a significant impact on FX reactions; when combined with low growth in China, exposure to China does not act as a significant buffer (triple interaction not statistically significant in Column (4)).

### Capital flow measures, international financial integration, and macroprudential policies
- Table 7 (Dependent Variable: 2-day Changes in Exchange Rate) findings:
  - Countries that imposed capital flow measures amidst large capital inflows (column 4) and countries with greater international financial integration (columns 5 and 6) experienced smaller exchange-rate depreciation.
  - Triple interaction (dummy × international financial integration × capital flow measures) is negative and statistically significant.
  - Dummy coefficients across specifications: 1.122***, 1.008***, 1.140***, 1.120***, 1.119***, 1.124***, 1.121*** (standard errors: (0.130), (0.127), (0.169), (0.156), (0.124), (0.164), (0.164))
  - Interactions reported:
    - Excess capital flows: -0.001, -0.002, -0.001 (standard errors: (0.002), (0.002), (0.002))
    - Capital flow measures: 0.004, 0.001, 0.032, -0.001, 0.020 (standard errors: (0.016), (0.018), (0.024), (0.017), (0.019))
    - Excess capital flows * CFM: -0.001** (0.000)
    - International financial integration: -0.002*, -0.002*, -0.002* (standard errors: (0.001), (0.001), (0.001))
    - International financial integration * CFM: -0.001* (0.000)
  - Observations: 272, 340, 255, 255, 357, 340, 340. R-squared: 0.236, 0.232, 0.244, 0.249, 0.230, 0.236, 0.237.
  - Notes: Capital flow measures index: 1 = tightening, -1 = loosening, 0 = no action; index cumulated from 2000-2013.
- Table 8 (Macroprudential policies and financial depth):
  - Countries that tightened macroprudential policies prior to 2013 experienced less FX depreciation.
  - However, interaction between event dummy, macroprudential, and financial depth (stock market capitalization) is positive — marginal benefits of macroprudential policies diminish with financial depth.
  - Dummy: 1.184***, 1.702***, 1.453*** (standard errors: (0.126), (0.149), (0.168))
  - Interactions:
    - Macro prudential: -0.023** (0.009), -0.052*** (0.006), -0.003 (0.023) across columns
    - Stock market cap/GDP: -0.008*** (0.001)
    - Macro prudential * Stock market cap: 0.001*** (0.000)
    - M2/GDP: -0.004*** (0.001)
    - Macro prudential * M2/GDP: -0.000 (0.000)
  - Variables without interaction:
    - Stock market cap/GDP: 0.018*** (0.005)
    - Macro prudential * Stock market cap: 0.001 (0.001)
    - M2/GDP: -0.118*** (0.035)
    - Macro prudential * M2/GDP: 0.002 (0.001)
  - Observations: 340, 340, 340. R-squared: 0.241, 0.271, 0.277.
- Interpretation: Tighter stance on both CFMs and macroprudential policies in the run-up to 2013-14 episodes helped mitigate negative market reactions; CFMs’ benefits increase with greater financial integration, while macroprudential benefits diminish with greater financial depth.

### Economic significance of estimates
- The estimates are economically significant and can reach up to 0.3 percentage points depreciation over two days (around 50 percent annualized) for increases in vulnerability of one standard deviation above average in areas listed (inflation, current account, fiscal balance, etc.).
- Example given: a one standard deviation higher vulnerability stemming from a 2.6 percentage point lower inflation rate is associated with (chart discussion in text; specific associated depreciation value referenced as part of the up to 0.3 percentage points figure).

*Source: IMF working paper content as provided in the supplied PDF excerpt.*

### 0.15 percentage point lower depreciation.

### _wp14109 - 0.15 percentage point lower depreciation.

### Effect on bond yields and equity prices
- Regression results for two-day changes in 10-year bond yields around events are qualitatively similar to those for exchange rates (Table 9).
- Countries with stronger fundamentals—larger current account balances and reserves; deeper financial markets; larger exposure to China; and tighter macroprudential policy stance in run-up to 2013 events—experienced smaller increases in bond yields during events of turmoil.
- The interaction coefficients on growth, inflation, fiscal balance, and financial integration are of the expected sign but not statistically significant.
- Example magnitude: "a one standard deviation higher reserves/GDP or lower inflation is associated with 4 basis points larger increase in yields over a two-day period."
- Evidence for equity prices is less compelling:
  - Interactions with most country characteristics are not statistically significant and often have unexpected signs (Table 10).
  - Exception: the interaction of the negative event dummy with excess capital inflows is negative and statistically significant—countries which received large inflows in the run-up to the tapering episodes experienced larger decline in equity prices.
- Interpretation: equity investors differentiated little across countries relative to bond investors; bond and FX markets appear to have had more short-term, carry-trade motivated participants who differentiated based on external financing needs.

### Which country characteristics matter more?
- A specification with interactions for all determinants (Table 11) yields weaker but still present evidence of market differentiation.
- Macro fundamentals that continue to matter include:
  - Reserves
  - Growth
  - Inflation
  - Change in current account
  - Foreign-exchange-denominated debt (stock)
- Specific statistically significant findings in Table 11 (interaction coefficients, two- and four-day horizons, selected entries preserved exactly):
  - dummy coefficients: 0.946*, 2.905***, 0.327**, 0.489**, -1.414, -2.342 (columns 1–6 respectively).
  - Interaction terms (selected examples):
    - CA/GDP: 0.014; 0.076**; 0.001; 0.003; -0.060; -0.101
    - Reserves/GDP: -0.004; -0.027***; -0.003*; -0.004; 0.024; 0.038
    - Change in CA/GDP 09-12: -0.003; -0.066**; -0.024***; -0.029***; 0.066; 0.140
    - Stock FX/GDP: 0.040***; -0.005; -0.007; -0.008; 0.060; 0.100*
    - Macro prudential: 0.011; -0.031**; -0.009***; -0.016***; -0.012; 0.057**
- Summary conclusion: better macro fundamentals, deeper financial markets, and tighter macroprudential policy stance in run up to 2013–14 episodes are associated with lower exchange rate depreciation and smaller increases in yields.

### Effect of overall vulnerability: Principal component analysis
- Method: normalize and standardize variables to lie between zero and one (higher values indicate lower vulnerability/greater resilience); take the first principal component from a large set of country-specific characteristics (fundamentals, financial depth, financial integration, emerging market growth, policy, exposure to China).
- Most resilient EMs based on the first principal component include Singapore, Taiwan, Malaysia, Korea, and Thailand.
- Table 12 main findings (2-day changes, coefficients preserved exactly):
  - Exchange rate: Dummy 1.032***; Dummy*Resilience -0.087***; Resilience 0.102; Observations 306; R-squared 0.249.
  - Government bond yields: Dummy 0.130***; Dummy*Resilience -0.017*; Resilience 0.054; Observations 306; R-squared 0.187.
  - Stock prices: Dummy -1.589***; Dummy*Resilience -0.008; Resilience 0.070; Observations 306; R-squared 0.212.
- Interpretation: the coefficient on Dummy*Resilience is negative and significant for exchange rate and bond yields—more resilient a country is, lower is the magnitude of depreciation and increase in bond yields. Resilience does not matter significantly around non-negative events. Resilience matters less for explaining stock price reactions.

### Robustness checks
- Alternative event measures:
  - Changes in expected future short-term rates around FOMC announcements (e.g., euro-dollar futures, OIS, Fed Fund futures) yield broadly similar results: countries with better fundamentals, greater financial depth, better growth prospects, greater exposure to China, and tighter macroprudential policies fared better.
- Table 13 (FX reaction to tightening using 2-day change in euro-dollar futures (4 years ahead), selected coefficients preserved exactly):
  - 2 day change in euro-dollar futures (4 years ahead): 4.112***, 5.239***, 5.949***, 3.584***, 3.849***, 5.570***, 6.037***, 5.464***, 5.012***, 3.611***, 4.329***, 7.600***, 5.030***, 5.048***, 4.667***, 4.450***, 4.064*** (columns 1–17).
  - Interaction examples:
    - Interaction of change in ED with CA/GDP: -0.120*** (0.042)
    - Reserves/GDP: -0.037*** (0.013)
    - Growth: -0.505*** (0.158)
    - Fiscal balance/GDP: -0.190*** (0.066)
    - Macro prudential: -0.096*** (0.032)
- Markets expectations of Fed asset purchases (New York Fed surveys):
  - Based on the June survey, markets expected Fed to purchase US$ 65 billion of QE assets in September (instead of US$ 85 bn in April).
  - Revision in expectations variable (RΔ) takes a value of 20 (US$ bn) for June when markets revised expectations downward by this amount, and zero otherwise.
- Table 14 (FX reaction to revision of market expectations of FED purchases, 2-day changes, selected coefficients preserved exactly):
  - Revision in market expectation: -0.00908***; -0.01149***; -0.00802**; -0.00410**; -0.00960***; -0.01370***; -0.01373***; -0.01417***; -0.01257***; -0.00948***; -0.01135***; -0.00850**; -0.00838; -0.01065***; -0.00816***; -0.01079***; -0.01272*** (columns 1–17).
  - Interactions with country characteristics (selected examples, exact values preserved):
    - Inflation interaction: -0.00140** [0.001]
    - Bank credit/GDP interaction: 0.00006** [0.000]
    - M3/GDP interaction: 0.00004* [0.000]
    - M2/GDP interaction: 0.00006*** [0.000]
    - Macro prudential interaction: -0.00053** [0.000]
    - Capital flow measures interaction: 0.00032* [0.000]
  - Main conclusion: when markets revised expectations of Fed purchases downwards (expecting tapering sooner), exchange rates depreciated; there is some evidence of differentiation by fundamentals (lower inflation, greater financial depth, tighter capital flow measures experienced lower depreciation).

### Conclusion (paper's main findings)
- The Fed’s plans and decision to taper were accompanied by significant market reaction in many EMs.
- Markets differentiated across countries during episodes of turmoil based on country characteristics, including macroeconomic fundamentals and economic and financial structures.
- Countries with better fundamentals and greater financial depth were less hit than others.
- Countries that had tighter macroprudential policies and had resorted to capital flow measures prior to May 2013 fared better.
- Strong trade ties with China can help dampen market reaction when no bad news emanates from China; when bad news from China coincides with the initial reason for turmoil in EMs, strong ties with China do not dampen market reactions.

*Source: IMF Staff estimates drawn from the provided document content.*

### REFERENCES

### _wp14109 - REFERENCES

### References
- Ahmed, Shahgil, Brahima Coulibaly, and Andrei Zlate, 2014, “Transmission of Shocks to Emerging Market Economies: How Important are Economic Fundamentals”, mimeo, Federal Reserve Board.
- Aizenman, Joshua, Mahir Banici, and Michael M. Hutchinson, 2014, “Transmission of Federal Reserve Tapering to Emerging Financial Markets”, NBER Working Paper No. 19980.
- Bernanke, B. S. and Kuttner, K. N. (2005), What Explains the Stock Market's Reaction to Federal Reserve Policy?. The Journal of Finance, 60: 1221–1257.
- Chen, Jiaqian, Tomasso Mancini-Griffoli, & Ratna Sahay, 2014. “US Monetary Policy Impact on Emerging Markets: Different This Time?,” forthcoming IMF Working Paper (Washington: International Monetary Fund).
- Eichengreen, Barry & Gupta, Poonam, 2014. "Tapering talk: the impact of expectations of reduced Federal Reserve security purchases on emerging markets," Policy Research Working Paper Series 6754, The World Bank.
- Gürkaynak, R.S., B Sack, E Swanson, 2005, “The sensitivity of long-term interest rates to economic news: Evidence and implications for macroeconomic models”, The American Economic Review 95 (1), 425-436.
- IMF, 2014a. “Financial Deepening in Emerging Markets”, Board Paper, forthcoming.
- IMF, 2014b, “Global Financial Stability Report, Chapter 2. How Do Changes in the Investor Base and Financial Deepening Affect Emerging Market Economies”?
- Rogoff, Kenneth, Aasim M. Husain, Ashoka Mody, Robin Brooks, and Nienke Oomes (2004), "Evolution and Performance of Exchange Rate Regimes," IMF Occasional Paper 229.
- Zhang, Longmei & Edda Zoli, 2014. “Leaning Against the Wind: Macroprudential Policy in Asia,” IMF Working Paper 14/22 (Washington: International Monetary Fund).

### Appendix I — Table A1. Summary Statistics (IMF Staff Estimates)
- Note: Source: IMF Staff Estimates.
- Variables (each entry lists Obs, Mean, Std. Dev., Min, Max, Source)
  - Change in FX (1-day before and after)
    - Obs: 357
    - Mean: 0.16
    - Std. Dev.: 0.99
    - Min: -3.10
    - Max: 3.65
    - Source: Bloomberg
  - Change in Yields (1-day before and after)
    - Obs: 357
    - Mean: 0.02
    - Std. Dev.: 0.15
    - Min: -0.80
    - Max: 0.84
    - Source: Bloomberg
  - Change in Equity (1-day before and after)
    - Obs: 357
    - Mean: -0.11
    - Std. Dev.: 1.63
    - Min: -5.96
    - Max: 6.26
    - Source: Bloomberg
  - Current Account/GDP
    - Obs: 357
    - Mean: 0.39
    - Std. Dev.: 5.80
    - Min: -7.22
    - Max: 18.60
    - Source: Bloomberg
  - Change in Current Account/GDP
    - Obs: 357
    - Mean: -1.33
    - Std. Dev.: 3.39
    - Min: -10.93
    - Max: 6.03
    - Source: Bloomberg
  - Reserves/GDP
    - Obs: 357
    - Mean: 30.78
    - Std. Dev.: 21.48
    - Min: 10.12
    - Max: 93.71
    - Source: Haver
  - Real GDP Growth
    - Obs: 357
    - Mean: 3.71
    - Std. Dev.: 2.35
    - Min: -2.70
    - Max: 19.10
    - Source: Bloomberg
  - Inflation
    - Obs: 357
    - Mean: 3.64
    - Std. Dev.: 2.56
    - Min: 0.20
    - Max: 10.56
    - Source: Bloomberg
  - Fiscal Balance/GDP
    - Obs: 357
    - Mean: -1.31
    - Std. Dev.: 3.93
    - Min: -10.78
    - Max: 7.92
    - Source: Haver
  - Corporate FX Debt/GDP
    - Obs: 357
    - Mean: 11.89
    - Std. Dev.: 6.71
    - Min: 2.56
    - Max: 28.24
    - Source: BIS
  - Bank Credit/GDP
    - Obs: 357
    - Mean: 74.69
    - Std. Dev.: 40.60
    - Min: 20.25
    - Max: 148.89
    - Source: IFS
  - M3/GDP
    - Obs: 289
    - Mean: 97.50
    - Std. Dev.: 54.64
    - Min: 43.87
    - Max: 242.11
    - Source: Haver
  - M2/GDP
    - Obs: 357
    - Mean: 79.01
    - Std. Dev.: 57.83
    - Min: 18.48
    - Max: 243.49
    - Source: Haver
  - Stock Market Capitalization/GDP
    - Obs: 357
    - Mean: 75.31
    - Std. Dev.: 53.27
    - Min: 8.16
    - Max: 223.70
    - Source: Bloomberg, Haver
  - Bid-Ask Spread of Government Bonds
    - Obs: 340
    - Mean: 0.0600
    - Std. Dev.: 0.056
    - Min: 0.002
    - Max: 0.345
    - Source: Bloomberg
  - One-year ahead Consensus Forecast of Real GDP Growth
    - Obs: 357
    - Mean: 4.27
    - Std. Dev.: 1.51
    - Min: 0.02
    - Max: 8.16
    - Source: Consensus Forecast
  - Current Consensus Forecast of Real GDP Growth
    - Obs: 357
    - Mean: 4.21
    - Std. Dev.: 1.87
    - Min: -1.80
    - Max: 9.20
    - Source: Consensus Forecast
  - Portfolio Asset and Liabilities/GDP
    - Obs: 357
    - Mean: 57.62
    - Std. Dev.: 69.30
    - Min: 7.02
    - Max: 336.87
    - Source: IFS
  - Cummulative Macroprudential Index
    - Obs: 340
    - Mean: 6.90
    - Std. Dev.: 8.67
    - Min: -1.00
    - Max: 40.00
    - Source: Zhang and Zoli (2014)
  - Excessive Capital Inflow
    - Obs: 272
    - Mean: 24.59
    - Std. Dev.: 39.34
    - Min: -52.36
    - Max: 107.25
    - Source: IMF 2014
  - Cumulative Capital Flow Measure Index
    - Obs: 340
    - Mean: 3.45
    - Std. Dev.: 5.13
    - Min: -1.00
    - Max: 17.00
    - Source: Zhang and Zoli (2014)
  - Export and Import to China/GDP
    - Obs: 340
    - Mean: 9.80
    - Std. Dev.: 8.47
    - Min: 1.60
    - Max: 33.21
    - Source: IFS

*Source: IMF Staff Estimates.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14109.pdf_