## _wp1429

## Source details

**Canonical URL:** [_wp1429](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp1429.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp1429.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp1429.pdf.json)

---

### Introduction: scope and main findings
- Objective: analyze influence of foreign investors on level and volatility of local currency government bond yields with a focus on the post-Lehman period.
- Sample period focus: post-crisis period (2009Q1 to 2013Q1).
- Key cross-country finding:
  - An increase in the share of foreign investors in EM local currency bond markets contributes to lower yields but higher yield volatility, the latter effect being mostly observed in countries with weak macroeconomic fundamentals.
- Poland case-study:
  - Share of foreign investors reached a peak at 37 percent in April 2013 from about 14 percent in early 2009.
  - Time-series (error-correction and GARCH) analysis for Poland: higher share of foreign holdings associated with lower yields and a positive, albeit smaller, association with yield volatility.

### Intuition and policy relevance
- Benefits of foreign investors:
  - Enable EMs to borrow abroad in their own currency, reducing “original sin”.
  - Catalyze development of local bond markets by diversifying investor base and increasing liquidity.
- Risks from larger foreign presence:
  - Greater susceptibility to sudden drying-up of capital flows from global risk aversion or rising global rates (example: Fed “tapering talk” on May 22).
  - Complicate macroeconomic management during large inflows (asset price or credit bubbles) and expose economies to flow reversals.
- Key policy implication:
  - Benefits from broadened international investor base are conditional on macroeconomic performance and buffers; EMs with weak fundamentals suffer most from high initial exposure to foreign investors and show higher ex-post yield volatility.

### Cross-country sample, variables, and econometric approach
- Countries included (12): Brazil, Czech Republic, Hungary, Indonesia, Republic of Korea, Malaysia, Mexico, Poland, Slovakia, South Africa, Thailand, and Turkey.
- Dependent variables:
  - Level: 5-year local currency government bond yields.
  - Volatility: quarterly standard deviation of weekly changes in 5-year yields (12 weeks per quarter).
- Controls:
  - Country-specific: central bank policy rates, inflation rate, change in real GDP growth rate; robustness: public debt-to-GDP, current account-to-GDP, foreign reserves-to-GDP, 2-year forward exchange rate vs. USD.
  - Global: VIX and U.S. federal funds rate.
- Panel specification: country and quarter fixed-effects; various specifications include pooled OLS, fixed-effects with AR(1) residuals, and IV estimation.

### Baseline empirical results — level of yields
- Main quantitative result:
  - A 10 percentage point increase in the share of foreign investors is associated with a reduction in yields of 70 to 90 bps.
- Selected coefficient examples (Table 1):
  - Foreign holdings ratio coefficients: Column (1) -0.0893***; Column (2) -0.0727***; Column (3) -0.0742***; Column (9) -0.0665***.
  - Policy rate: Column (1) 0.647***; Column (9) 0.441***.
  - International reserves-to-GDP: Column (8) -0.0891*** (example: a 1 percentage point increase in foreign exchange reserves-to-GDP lowers yields by about 9 basis points on average).
  - Forward exchange rate: Column (9) 0.00130***.
  - U.S. federal funds rate: Column (3) -2.478**; Column (8) -2.795**.
- Observations and fit:
  - Observations across specifications vary: 191, 186, 176, 170, 154, 113.
  - R-squared examples: 0.224 up to 0.613.

### Baseline empirical results — volatility of yields
- Pooled OLS: negative association between foreign holdings ratio and yield volatility (weaker significance after controls).
- With country fixed-effects: foreign holdings ratio generally not associated with a decline in yield volatility; point estimates often negative but not statistically significant.
- Robust correlates of volatility:
  - VIX is robustly and significantly positively associated with yield volatility (e.g., Table 2: VIX, ln Column (2) 1.215**; Column (4) 1.304***).
  - Volatility of forward exchange rate positively and significantly associated with yield volatility (Table 2 Column (5) coefficient 0.799***).
- Caveat: fixed-effects specifications do not fully rule out endogeneity or conditional effects tied to macroeconomic fundamentals.

### Identification and robustness checks
- Serial correlation:
  - Panel fixed-effects with AR(1) residuals (Baltagi and Wu, 1999) yields negative and significant foreign holding coefficient within same magnitude range.
- Outliers:
  - Excluding influential observations (residuals > two standard deviations) does not overturn negative association between foreign holdings and yields.
- Endogeneity addressed via IVs:
  - Instruments:
    - Second and third lags of the foreign holdings variable.
    - Time-varying predictions from regressions of foreign holdings on financial remoteness (ln great-circle distance to closest offshore financial center), possibly including quadratic terms.
  - First-stage and diagnostics:
    - First-stage: foreign holdings, t-2 coefficients around 0.95–1.01*** with strong t-stats.
    - Kleibergen Paap and Cragg-Donald F-statistics examples: 360.70; 425.71; 59.5; 70.8; 37.8.
    - Hansen OID test p-values reported as satisfactory in examples (e.g., 0.563).
  - IV results:
    - IV estimates support a negative and statistically significant effect of foreign holdings on the level of yields; IV second-stage examples include -0.0389** (Column (4)) and -0.0387** (Column (5)).
    - IV estimates for volatility indicate a positive and significant association between higher foreign investor dependency and yield volatility (Table 4: second-stage foreign holding ratio Column (1) 0.0389**; Column (2) 0.0587*; Column (3) 0.116**).
    - Financial remoteness-based IVs produce higher estimated impacts.

### Non-linear effects and thresholds (macro fundamentals)
- Interaction specifications between foreign holdings and fundamentals include gross government debt, total external debt, current account balance, international reserves, and VIX.
- Level of yields:
  - No statistically significant non-linear effect of foreign holdings on level of yields except interactions with gross public debt and total external debt ratios (e.g., Foreign holdings * Public gross debt ratio 0.00129***).
  - Caveat: computed debt-to-GDP thresholds are out-of-sample thresholds and should be interpreted with caution.
- Yield volatility:
  - Country fundamentals condition the effect of foreign holdings on volatility:
    - Positive association between foreign share and yield volatility increases with the level of total external debt-to-GDP (Foreign holdings * Gross external debt-to-GDP 0.000465**).
    - Positive association decreases with stronger balance of payments (Foreign holdings * Current account balance-to-GDP -0.00212*; Foreign holdings * International reserves-to-GDP -0.00179**).
    - Effect does not appear to depend on global risk aversion (VIX).
  - Threshold example:
    - Total gross external debt above 98 percent of GDP is conducive to an increase in yield volatility in response to higher foreign holdings.
  - Interpretation:
    - Countries with good fundamentals: foreign inflows reduce yields and may cause limited adjustment volatility.
    - Countries with weak fundamentals: large foreign inflows raise risk of reversals and lead to more pronounced yield volatility.

### Poland case study — methods and main results
- Context:
  - Foreign investor share rose from about 14 percent in early 2009 to 37 percent in April 2013.
  - Time-series analysis period: end–2004 to August 2013; subsample end–2008 to August 2013.
  - Data frequencies: weekly, daily, monthly (foreign holdings only monthly).
- Estimation strategy (three-step):
  - Step 1: Error-correction model (weekly) to link yields to reference rate, FRA outlooks, log(VIX), log(CDS), FED balance sheet.
  - Step 2: GARCH (daily) to study role of global financial factors on level and volatility.
  - Step 3: GARCH (monthly) including share of foreign holdings to link mean and variance of yields to foreign holdings.
- Error-correction (weekly) results:
  - Reference rate and interest rate outlook explain 87.9 percent to 88 percent of variation in the yield curve; adding VIX and CDS raises explanatory power to 95.3 percent.
  - Short-run: Fed balance sheet (one-week lag) associated with higher yields; 4-week lag may depress yields. Higher VIX and CDS increase yields. Forward exchange appreciation associated with lower yields. Error-correction term significant and negative.
- GARCH (daily) results:
  - Mean: changes in reference rate and outlooks affect yields; ∆log(VIX) and ∆log(CDS) positively influence yield changes; US5y positive; forward appreciation lowers yield.
  - Variance: bond yield volatility increases with VIX; VIX effect on volatility increases after 2008–09 crisis. Results robust across GARCH specs.
- GARCH (monthly with foreign holdings) results:
  - Full sample (Dec 2004–Jul 2013): greater foreign ownership reduced mean yields; impact on yield volatility insignificant (consistent with Peiris (2010)).
  - Subsample (Dec 2008–Jul 2013): greater foreign ownership associated with lower yields and simultaneously with higher yield volatility.
  - Risk-scenarios: Poland vulnerable to sudden-stop of inflows, passive sales, active sales due to higher initial foreign exposure and limited domestic bank absorption capacity.
- Selected Poland coefficient highlights:
  - Long-run regressions: reference rate coefficients around 0.54*** to 1.05***; log(VIX) and log(CDS) positive and sometimes significant; R-squared up to 0.973.
  - GARCH including foreign ownership (Table 10): foreign ownership in mean equation examples -0.021 and -0.015***; variance effects examples -0.000041 and 0.00040*** (subsample).

### Summary implications and policy recommendations
- Empirical summary:
  - Foreign inflows into EM local-currency bond markets lower yield levels across EMs.
  - Higher foreign participation increases yield volatility, especially in countries with weak external or fiscal buffers (higher external debt, weaker current account, lower reserves).
  - Global factors (VIX, US yields) are important drivers of both levels and volatility; the influence of global factors on volatility rose after the 2008–09 crisis.
- Policy recommendations:
  - Continue improving macroeconomic fundamentals and build policy buffers.
  - Adopt prudent fiscal policies and measures to encourage national savings.
  - Manage public debt prudently, attract long-term foreign investors, and create liquid fiscal buffers to deal with uncertainties.
  - Monitor foreign investor exposure and strengthen external and fiscal buffers to mitigate volatility risks associated with greater foreign participation.

### Key statistics and descriptive data (selected from Appendix A and tables)
- Summary sample statistics (quarterly data):
  - 5-year local currency government bond yield: Obs 204; Mean 5.99; Std. Dev. 3.04; Min 0.96; Max 16.96.
  - Quarterly yield volatility: Obs 204; Mean 2.72; Std. Dev. 1.72; Min 0; Max 10.44.
  - Foreign holdings (in percent): Obs 191; Mean 21.10; Std. Dev. 10.99; Min 2.60; Max 46.90.
  - Policy rate (in percent): Obs 199; Mean 4.35; Std. Dev. 2.70; Min 0.05; Max 12.25.
  - VIX CBOE, ln: Obs 204; Mean 3.14; Std. Dev. 0.33; Min 2.72; Max 3.80.
  - U.S. Federal funds rate: Obs 204; Mean 0.15; Std. Dev. 0.05; Min 0.07; Max 0.22.
  - External debt (in percent of GDP): Obs 204; Mean 46.07; Std. Dev. 33.30; Min 9.91; Max 149.52.
  - Gross public debt (in percent of GDP): Obs 198; Mean 44.72; Std. Dev. 17.28; Min 22.41; Max 100.86.
  - Foreign exchange reserves (in percent of GDP): Obs 187; Mean 23.92; Std. Dev. 13.07; Min 8.35; Max 55.27.
- Thresholds and notable cutoffs:
  - Total gross external debt above 98 percent of GDP associated with higher yield volatility response to foreign holdings.
  - Example conditional values mentioned in the content: 1.5 percent of GDP or foreign exchange reserves close to 12 percent of GDP (in relation to yield volatility contexts reported in the source).

*Source: IMF working paper content (post-Lehman sample 2009Q1–2013Q1) as provided in the content unit.*

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

### _wp1429 - References .............................................................................................................

### Introduction: scope and main findings
- Objective: analyze influence of foreign investors on level and volatility of local currency government bond yields with a focus on the post-Lehman period.
- Sample period focus: post-crisis period (2009Q1 to 2013Q1).
- Key cross-country finding:
  - An increase in the share of foreign investors in EM local currency bond markets contributes to lower yields but higher yield volatility, the latter effect being mostly observed in countries with weak macroeconomic fundamentals.
- Country case-study: Poland
  - Share of foreign investors reached a peak at 37 percent in April 2013 from about 14 percent in early 2009.
  - Time-series (error-correction and GARCH) analysis for Poland: higher share of foreign holdings associated with lower yields and a positive, albeit smaller, association with yield volatility.

### Intuition and policy relevance
- Positive roles of foreign investors:
  - Enable EMs to borrow abroad in their own currency, reducing “original sin”.
  - Can catalyze development of local bond markets by diversifying investor base and increasing liquidity.
- Risks from larger foreign presence:
  - Greater susceptibility to sudden drying-up of capital flows from global risk aversion or rising global rates (example: Fed “tapering talk” on May 22).
  - Potential to complicate macroeconomic management during large inflows (asset price or credit bubbles) and expose economies to flow reversals.

### Analytical challenges highlighted
- Endogeneity: direction of causality between foreign demand and yields likely goes both ways.
- Unobservable forward-looking determinants of long-term interest rates (e.g., expectations of long-run inflation) complicate identification.
- Simple correlations can be misleading; negative endogeneity bias (foreign investors avoid high-volatility markets) must be corrected to infer causal effects.

### Cross-country estimates: sample and approach
- Countries included (12): Brazil, Czech Republic, Hungary, Indonesia, Republic of Korea, Malaysia, Mexico, Poland, Slovakia, South Africa, Thailand, and Turkey.
- Dependent variables:
  - Level: 5-year local currency government bond yields.
  - Volatility: quarterly standard deviation of weekly changes in 5-year yields (12 weeks per quarter).
- Panel specification: country and quarter fixed-effects, controls for country-specific and global factors.
- Country-specific controls include: central bank policy rates, inflation rate, change in real GDP growth rate; additional robustness controls: public debt-to-GDP, current account-to-GDP, foreign reserves-to-GDP, 2-year forward exchange rate vs. USD.
- Global controls: VIX (Chicago Board Options Exchange Volatility Index) and U.S. federal funds rate.

### Baseline empirical results (level of yields)
- Main result: foreign participation has a statistically significant and negative effect on the level of 5-year local bond yields.
  - Magnitude: a 10 percentage point increase in the share of foreign investors is associated with a reduction in yields of 70 to 90 bps.
- Other significant correlates:
  - Country policy rate: strong and robust positive effect on yields.
  - International reserves-to-GDP: statistically significant and negative (example: a 1 percentage point increase in foreign exchange reserves-to-GDP lowers yields by about 9 basis points on average).
  - U.S. federal funds rate and current account balance: also play significant roles.
  - Expectations of currency depreciation (2-year forward): positively associated with higher bond yields.

### Baseline empirical results (volatility of yields)
- Pooled OLS (no country fixed-effects): negative association between foreign holdings ratio and yield volatility (weaker significance after controls).
- With country fixed-effects: foreign holdings ratio generally not associated with a decline in yield volatility; point estimates often negative but not statistically significant.
- Robust finding: VIX is robustly and significantly positively associated with yield volatility across specifications.
- Volatility of forward exchange rate: positively and significantly associated with yield volatility.
- Caveat: fixed-effects specifications do not fully rule out endogeneity or conditional effects tied to macroeconomic fundamentals.

### Robustness checks and identification strategies
- Serial correlation:
  - Panel fixed-effects model with residuals following AR(1) (Baltagi and Wu, 1999) yields negative and significant foreign holding coefficient within same magnitude range.
- Outliers:
  - Excluding influential observations (residuals > two standard deviations) does not overturn negative association between foreign holdings and yields.
- Endogeneity addressed via instrumental variables:
  - Instruments used:
    - Second and third lags of the foreign holdings variable.
    - Time-varying predictions from regressions of foreign holdings on financial remoteness (distance to offshore financial centers), possibly including quadratic term.
  - Financial remoteness instrument:
    - Uses natural logarithm of great-circle distance to closest offshore financial center (40 OFCs per Rose and Spiegel) with quarter-by-quarter OLS to generate time-varying predicted foreign holdings.
    - Rationale: countries closer to financial centers are more financially integrated and attract higher foreign shares; geography-based measure has plausible exogeneity.
  - Results:
    - IV estimates (lags and financial remoteness approaches) support a negative and statistically significant effect of foreign holdings on the level of yields.
    - First-stage F-statistics and Hansen tests indicate instrument relevance and orthogonality satisfactory (authors report diagnostic tests consistent).
- Volatility model IV results:
  - Regardless of IV method (lags or financial remoteness), estimates indicate a positive and significant association between higher foreign investor dependency and yield volatility.
  - Financial remoteness-based IVs produce higher estimated impacts.

### Non-linear effects: macroeconomic fundamentals and thresholds
- Specification: interaction terms between foreign holdings and macro fundamentals (gross government debt, total external debt, current account balance, international reserves, and global risk aversion).
- Level of yields:
  - No statistically significant non-linear effect of foreign holdings on level of yields except for gross public debt and total external debt ratios.
  - Noted caveat: computed debt-to-GDP thresholds are out-of-sample thresholds; results may be statistical artifacts and should be interpreted with caution.
- Yield volatility:
  - Country fundamentals matter for the relationship between foreign holdings and yield volatility:
    - The positive association between foreign share and yield volatility increases with the level of total external debt-to-GDP.
    - The positive association decreases with stronger balance of payment position (lower current account deficits and higher foreign exchange reserves).
    - Effect of foreign holdings does not seem to depend on degree of global risk aversion (VIX).
  - Threshold computations:
    - Estimates allow computation of the conditional-variable levels at which the marginal effect of foreign holdings on volatility is neutralized (details in Table 6).

### Poland case-study (high-frequency time-series analysis)
- Methods: error-correction model and GARCH framework using higher-frequency data.
- Result: increase in foreign holdings associated with lower yields and a positive association with yield volatility in Poland.
- Contrasts with pre-crisis Peiris (2010) finding (2000–09) which did not find a positive association between foreign holdings and yield volatility in Poland.
- Possible explanation: Poland’s role as a proxy for investors seeking exposure to Central and Eastern Europe.

### Summary implications
- Foreign inflows into EM local-currency bond markets:
  - Deliver immediate benefits in lowering yields across EMs.
  - Increase yield volatility, with the volatility effect amplified in countries with weak external or fiscal buffers (higher external debt, weaker current account, lower reserves).
- Policy relevance:
  - Benefits from broadened international investor base are conditional on macroeconomic performance and buffers.
  - Monitoring and strengthening external and fiscal buffers can mitigate volatility risks associated with greater foreign participation.
  - Identification strategies (lags, geographic financial remoteness) are useful to address endogeneity in empirical assessment.

*Source: IMF working paper content (post-Lehman, sample 2009Q1–2013Q1) as provided in the content unit.*

### 1.5 percent of GDP or foreign exchange reserves close to 12 percent of GDP, yield volatility

### _wp1429 - 1.5 percent of GDP or foreign exchange reserves close to 12 percent of GDP, yield volatility

### Key cross-country findings on foreign holdings and yield volatility
- An increase in the share of foreign holdings in a given country is associated with a decline in the country’s yield volatility when the country posts a current account surplus.
- For countries with variables below certain in-sample thresholds, the resulting volatility of yields is higher.
- Total gross external debt above 98 percent of GDP is conducive to an increase in yield volatility in response to higher foreign holdings.
- The differentiation implies two volatility channels:
  - Countries with good fundamentals: foreign inflows reduce yields and may cause some downward adjustment volatility.
  - Countries that later suffer deterioration in fundamentals: large foreign inflows increase the risk of flow reversals or sudden-stops and lead to more pronounced yield volatility.
- Policy-relevant implication: Emerging Markets (EMs) with weak fundamentals suffer most from high initial exposure to foreign investors in domestic bond markets and show higher ex-post yield volatility.

### Poland case study — overview and motivation
- Poland experienced a large increase in foreign investor share in its domestic bond market: from about 14 percent in early 2009 to a peak at 37 percent in April 2013.
- Associated development: decline in bond yield levels over the same period.
- Time-series analysis period: end–2004 to August 2013. Subsample period: end–2008 to August 2013 (post large increase in foreign ownership).
- Data frequency used: weekly, daily, and monthly (foreign holdings only available at monthly frequency).

### Estimation strategy (three-step sequential procedure)
- Step 1: Error-correction model with weekly data to assess relationship between domestic bond yields and traditional indicators (expected future interest rates, global factors such as global risk aversion and the U.S. Federal Reserve balance sheet).
- Step 2: GARCH specification with daily data to assess the role of global financial factors on the level and volatility of yields in Poland.
- Step 3: GARCH model with monthly data including the share of foreign holdings to link both the mean and variance of yields directly to foreign holdings.

### Error-correction model (weekly data) — setup and results
- Variable focus: 5-year local currency bond yields (ݕ5ܮܲ
௧
) with weekly data from December 31, 2004 to August 30, 2013.
- Long-run regressors: week-end NBP reference rate (ܴܲܮܲ
௧
), FRA (3m,6m) and FRA (21m,24m) rate outlooks, log(VIX), log(5-year sovereign CDS for Poland).
- Long-run explanatory power:
  - Reference rate and interest rate outlook explain 87.9 percent of variation in the yield curve (Column 3 of Table 8).
  - Reference rate and long-run rate outlook account for 88 percent of variation; reference rate and short-run rate outlook account for 67.1 percent of variation (Table 7).
  - Adding VIX and CDS increases explanatory power to 95.3 percent of variation.
- Short-run findings:
  - Federal Reserve balance sheet (one-week lag) is associated with higher Polish government bond yields; a more lagged larger balance sheet (4-week lag) seems to depress the yield.
  - Changes in reference rate and in short-run and long-run interest rate outlook influence short-term yields as expected.
  - Higher VIX and higher CDS associated with increases in bond yields.
  - A more appreciated forward exchange rate (2 year forward USD/PLN) is associated with lower bond yield.
  - Error-correction term is significant with expected negative sign, confirming stable adjustment dynamics.
- Subsample (December 31, 2008 to August 30, 2013) difference:
  - CDS coefficient becomes larger; VIX, Fed balance sheet, and forward exchange rate direction may cease to be significant, suggesting CDS dominated after the 2008–09 financial crisis.

### GARCH model using daily data — setup and results
- Daily sample: December 31, 2004 to August 30, 2013.
- Mean equation regressors include: changes in reference rate and rate outlooks, ∆log(VIX), ∆log(CDS), US5y (five year US government bond yield), and forward exchange rate appreciation ∆ (2-year forward USD/PLN).
- Mean equation findings:
  - Changes in reference rate and interest rate outlook influence the yield.
  - ∆log(VIX) and ∆log(CDS) positively influence yield changes (VIX coefficient not always significant).
  - US5y is significant with expected positive sign; in the subsample (December 31, 2008 to August 30, 2013) the US bond yield coefficient becomes larger.
  - More appreciated forward exchange rate associated with lower bond yield.
- Variance equation findings:
  - Volatility of bond yields increases as VIX increases, confirming global risk aversion impacts volatility.
  - Influence of VIX on volatility increased after the 2008–09 financial crisis.
  - Results robust across different GARCH specifications and error distributions.
- Interpretation: volatility increasingly influenced by global factors linked to higher foreign holdings of local currency government bonds.

### GARCH model using monthly foreign holdings data — setup and results
- Model: GARCH(1,1) with monthly foreign holdings share (ܨ
௧
) included directly.
- Monthly sample: December 2004 to July 2013; subsample: December 2008 to July 2013.
- Full-sample results (December 2004 to July 2013):
  - Greater foreign ownership reduced the mean of the yields.
  - Impact on yield volatility is insignificant (consistent with Peiris (2010)).
- Subsample results (December 2008 to July 2013):
  - Greater foreign ownership associated with lower yields and simultaneously with higher yield volatility.
  - This reflects a mixed role of foreign investors and aligns with cross-country evidence and simulation results in Arslanalp and Tsuda (2013).
  - Risk-scenarios suggest Poland would be significantly hit by changes in foreign investor behavior (sudden-stop of fund inflows, passive sales, active sales) due to combination of higher initial exposure to foreign investor base and limited domestic bank absorption capacity.

### Conclusion and policy recommendations
- Main conclusions:
  - Foreign investor participation in EM local currency government bond markets lowers yield levels but increases susceptibility to market sentiment and yield volatility.
  - Benefit of lower yields from foreign investors is universal across EMs regardless of macroeconomic fundamentals when global liquidity is ample.
  - When global liquidity tightens, investors differentiate across EMs, and EMs with weak fundamentals suffer more from higher yield volatility if exposed to large foreign holdings.
- Policy recommendations:
  - Continue improving macroeconomic fundamentals and build policy buffers.
  - Adopt prudent fiscal policies and other measures to encourage national savings.
  - Manage public debt prudently, attract long-term foreign investors, and create liquid fiscal buffers to deal with uncertainties.

*Source: _wp1429 - 1.5 percent of GDP or foreign exchange reserves close to 12 percent of GDP, yield volatility (PDF chapter/section).*

### References

### References and Appendixes (Content Unit: _wp1429 - References)

### Key Literature Cited
- Andritzky, M.J.R., 2012. Government Bonds and Their Investors: What Are the Facts and Do They Matter? International Monetary Fund.
- Arslanalp, Serkan, Tsuda, Takahiro, 2013. Tracking the Global Demand for Emerging Market Sovereign Debt (IMF Working Paper, Forthcoming). International Monetary Fund.
- Baldacci, E., Gupta, S., Mati, A., 2011. Political and Fiscal Risk Determinants of Sovereign Spreads in Emerging Markets. Review of Development Economics 15, 251–263.
- Baltagi, B.H., Wu, P.X., 1999. Unequally Spaced Panel Data Regressions with AR(1) Disturbances. Econometric Theory 15, 814–823.
- Beltran, D.O., Kretchmer, M., Marquez, J., Thomas, C.P., 2013. Foreign holdings of U.S. Treasuries and U.S. Treasury yields. Journal of International Money and Finance 32, 1120–1143.
- Calvo, G.A., Izquierdo, A., Talvi, E., 2006. Sudden Stops and Phoenix Miracles in Emerging Markets. The American Economic Review 96, 405–410.
- Comelli, F., 2012. Emerging market sovereign bond spreads: Estimation and back-testing. Emerging Markets Review 13, 598–625.
- Coval, J., J. Moskowitz, T., 2001. The Geography of Investment: Informed Trading and Asset Prices. Journal of Political Economy 109, 811–841.
- Coval, J.D., Moskowitz, T.J., 1999. Home Bias at Home: Local Equity Preference in Domestic Portfolios. The Journal of Finance 54, 2045–2073.
- Csonto, B., Ivaschenko, I.V., 2013. Determinants of Sovereign Bond Spreads in Emerging Markets: Local Fundamentals and Global Factors vs. Ever-Changing Misalignments (IMF Working Paper No. 13/164). International Monetary Fund.
- International Monetary Fund, 2012. Global Financial Stability Report, October 2012: Restoring Confidence and Progressing on Reforms. International Monetary Fund.
- Jaramillo, L., Weber, A., 2012. Bond Yields in Emerging Economies: It Matters What State You Are In (IMF Working Paper No. 12/198). International Monetary Fund.
- Malloy, C.J., 2005. The Geography of Equity Analysis. The Journal of Finance 60, 719–755.
- Peiris, S.J., 2010. Foreign Participation in Emerging Markets’ Local Currency Bond Markets (IMF Working Paper No. 10/88). International Monetary Fund.
- Petersen, M.A., Rajan, R.G., 2002. Does Distance Still Matter? The Information Revolution in Small Business Lending. The Journal of Finance 57, 2533–2570.
- Rose, A., Spiegel, M., 2007. Offshore Financial Centres: Parasites or Symbionts? Economic Journal 117, 1310–1335.
- Rose, A.K., Spiegel, M.M., 2009. International financial remoteness and macroeconomic volatility. Journal of Development Economics 89, 250–257.
- Wu, T., 2006. Macro Factors and the Affine Term Structure of Interest Rates. Journal of Money, Credit and Banking 38, 1847–1875.

### Appendix A — Descriptive Statistics: Quarterly Data (Table A1)
- Observations and sample statistics (selected variables):
  - 5-year local currency government bond yield: Obs 204; Mean 5.99; Std. Dev. 3.04; Min 0.96; Max 16.96
  - Quarterly yield volatility: Obs 204; Mean 2.72; Std. Dev. 1.72; Min 0; Max 10.44
  - Foreign holdings (in percent): Obs 191; Mean 21.10; Std. Dev. 10.99; Min 2.60; Max 46.90
  - Policy rate (in percent): Obs 199; Mean 4.35; Std. Dev. 2.70; Min 0.05; Max 12.25
  - ln (100+Inflation rate): Obs 204; Mean 4.64; Std. Dev. 0.03; Min 4.57; Max 4.76
  - Real GDP growth (in percent): Obs 192; Mean 0.75; Std. Dev. 10.58; Min -48.45; Max 93.14
  - VIX CBOE, ln: Obs 204; Mean 3.14; Std. Dev. 0.33; Min 2.72; Max 3.80
  - U.S. Federal funds rate: Obs 204; Mean 0.15; Std. Dev. 0.05; Min 0.07; Max 0.22
  - Current account balance (in percent of GDP): Obs 204; Mean -0.23; Std. Dev. 5.01; Min -11.60; Max 16.70
  - Overall fiscal balance (in percent of GDP): Obs 204; Mean -3.00; Std. Dev. 3.61; Min -13.28; Max 7.42
  - External debt (in percent of GDP): Obs 204; Mean 46.07; Std. Dev. 33.30; Min 9.91; Max 149.52
  - Gross public debt (in percent of GDP): Obs 198; Mean 44.72; Std. Dev. 17.28; Min 22.41; Max 100.86
  - Foreign exchange reserves (in percent of GDP): Obs 187; Mean 23.92; Std. Dev. 13.07; Min 8.35; Max 55.27

### Appendix B — Data Sources and Variable Definitions (Table A2)
- Cross-country analysis data sources include DataStream, Haver analytics, Asianbondonline, IMF World Economic Outlook Database.
- Key variable definitions preserve original phrasing:
  - 5-year local currency government bond yield: 5-year nominal yield on treasury securities in local currency (in percent)
  - VIX CBOE: Chicago Board Options Exchange (CBOE) volatility index
  - Forward exchange rate: Local currency to US dollar 2 year forward exchange rate
  - Foreign holdings ratio: Share of foreign investors’ holdings of domestic government bonds as a percentage of total outstanding amount of local currency bonds (in percent)
  - Inflation rate: Consumer Prices, period average, seasonally adjusted, quarter-over-quarter percent change, annualized
  - Real GDP growth rate: Gross domestic product, constant prices, National Currency, percent change
  - Current account balance ratio: Current account balance as a percentage of nominal GDP
  - Overall fiscal balance ratio: central or government fiscal balance as a percentage of nominal GDP (depending on the country)
  - External debt ratio, Public debt ratio, Foreign exchange reserves ratio: described and sourced to Haver analytics

- Poland-specific series (case study) sourced from DataStream include: 5-year local currency government bond yield; NBP reference rate; FRA3X6; FRA21X24; CDS 5-year sovereign CDS for Poland; VIX CBOE; US5y; FED Base money; USPLN 2 year forward.

### Figures (visual summaries referenced)
- Figure 1: Non-Resident Holdings Ratios in Emerging Markets, 2007–13 — foreign holdings presented as percent of total (series across quarters 2007q1–2013q1).
- Figure 2: Panel correlation between non-resident holdings, 5-year bond yields, and yield volatility (Quarterly Data 2009–13). Note: y-axis represents 5-year local currency government bond yields (left panel) and quarterly standard deviation of weekly changes (right panel).
- Figure 3: Poland — Time series of Five-year local currency government bond yields (Percent) and Foreign Ownership of Local Currency Government Bonds (in percent) for dates 12/31/04–6/30/13.

### Key Empirical Results — Cross-Country Panel Regressions (Selected Tables)
- Table 1: Baseline estimates (Dependent: Emerging Markets’ Local Currency 5Y Bond Yield; Period 2009q1-2013q1)
  - Foreign holdings ratio coefficients (selected columns):
    - Column (1): -0.0893*** (t-statistic [-7.170])
    - Column (2): -0.0727*** ([-8.206])
    - Column (3): -0.0742*** ([-8.316])
    - Column (9): -0.0665*** ([-5.744])
  - Policy rate coefficient examples:
    - Column (1): 0.647*** ([9.248])
    - Column (9): 0.441*** ([4.277])
  - U.S. Federal Fund rate examples:
    - Column (3): -2.478** ([-2.138])
    - Column (8): -2.795** ([-2.117])
  - Forward exchange rate: Column (9) coefficient 0.00130*** ([4.597])
  - International reserves-to-GDP: Column (8) coefficient -0.0891*** ([-2.874])
  - Observations across specifications vary: 191, 186, 176, 170, 154, 113.
  - R-squared examples: 0.224 up to 0.613 depending on specification.
  - Note: Country fixed-effects included in all reported specifications. Significance: *** p<0.01, ** p<0.05, * p<0.1.

- Table 2: Panel fixed-effects estimates (Dependent: Quarterly Yield volatility; Period 2009q1-2013q1)
  - Foreign holdings ratio:
    - Column (1): -0.0284** ([-2.415])
    - Column (2): -0.0233* ([-1.763])
    - Later columns show coefficients less significant, up to 0.0309 ([1.584]).
  - VIX, ln coefficients:
    - Column (2): 1.215** ([3.001])
    - Column (4): 1.304*** ([3.441])
  - Forward exchange rate volatility: Column (5) coefficient 0.799*** ([3.463])
  - Observations: commonly 191; R-squared range: 0.031 to 0.183.
  - Country fixed-effects: some specifications include them.

- Table 3: Robustness checks for yield level (IV and other checks)
  - Foreign holdings ratio:
    - IV Second-stage examples: Column (4) -0.0389** ([-2.350]); Column (5) -0.0387** ([-2.081])
    - Non-IV columns show -0.0683***, -0.0681***, -0.0912*** with t-stats provided.
  - First-stage instruments:
    - Foreign holdings, t-2: 1.012*** ([8.202])
  - Kleibergen Paap F-Stat. examples: 360.70; 59.5; 37.8
  - Cragg-Donald F-Stat. examples: 425.71; 70.8; 37.7
  - R-squared examples: 0.557; 0.511; 0.496

- Table 4: IV estimates for yield volatility (panel fixed-effects)
  - Second-stage foreign holding ratio:
    - Column (1): 0.0389** ([1.972])
    - Column (2): 0.0587* ([1.932])
    - Column (3): 0.116** ([2.175])
  - First-stage:
    - Foreign holdings, t-2: 0.955*** ([8.112])
  - Cragg-Donald F-Stat. examples: 399.6; 62.2; 28.2
  - Hansen OID test P-value example: 0.563

- Table 5: Non-linear effects on yield level
  - Baseline foreign holdings ratio coefficients:
    - Column (1): -0.0804*** ([-8.473])
    - Column (2): -0.107*** ([-6.586])
    - Column (3): -0.139*** ([-6.955])
    - Column (4): -0.0185 ([-0.375])
  - Interaction examples:
    - Foreign holdings * Gross external debt ratio: 0.000509* ([1.968])
    - Foreign holdings * Public gross debt ratio: 0.00129*** ([3.382])
  - Current account balance-to-GDP: Column (1) -0.0704** ([-2.066]); others similar.
  - Observations: 170–180; R-squared: 0.520–0.582.

- Table 6: Non-linear effects on yield volatility (panel fixed-effects)
  - Foreign holding ratio examples:
    - Column (4): -0.0457** ([-2.455])
  - Significant interactions:
    - Foreign holdings * Current account balance-to-GDP: -0.00212* ([-1.855])
    - Foreign holdings * International reserves-to-GDP: -0.00179** ([-2.601])
    - Foreign holdings * Gross external debt-to-GDP: 0.000465** ([2.488])
  - VIX, ln: coefficients often positive and significant, e.g., Column (1) 1.292*** ([3.774])
  - Joint significance p-values for non-linearity vary (examples: 0.221; 0.004; 0.104; 0.059; 0.165)
  - Threshold values of conditional variables and sample sizes provided in table.

### Poland Case Study — Yield Dynamics and Foreign Ownership (Selected Results)
- Table 7: Poland — Long-run dynamics (Dependent: Local Currency 5Y Bond Yield; Sample 2004/12-2013/8)
  - Reference rate coefficients: 0.54***, 1.05***, 1.03***, 1.03***, 0.89*** (Std. errors reported in brackets)
  - Long-run rate outlook: 0.85***, 0.80***, 0.82***, 0.82*** (Std. errors in brackets)
  - Log(VIX): 0.39***, 0.37*** (when included)
  - Log(CDS): 0.063***, 0.41*** (when included)
  - R-squared examples: 0.671; 0.877; 0.879; 0.953; 0.973
  - Number of observations: 453 (most columns), 245 (column 5).

- Table 8: Poland — Short-run dynamics (Dependent: ∆ Local Currency 5Y Bond Yield)
  - ∆Reference rate: 0.77***; 0.62***; 0.61*** (Std. errors in brackets)
  - ∆Short-run rate outlook: 0.23***; 0.15**; 0.14**
  - ∆Long-run rate outlook: 0.53***; 0.50***; 0.50***
  - ∆Log(CDS): 0.10**; 0.24***; 0.31***
  - ∆Log(Fed balance sheet) (one-week lag): 0.79*** and 0.14 (one specification)
  - Error correction term (one lag): -0.11***; -0.21***; -0.20***
  - R-squared: 0.743; 0.645; 0.641
  - Observations: 448; 244; 244

- Table 9: Poland — GARCH for ∆ Local Currency 5Y Bond Yield
  - Mean equation highlights:
    - ∆Reference rate: 0.69*** and 0.53***
    - ∆Short-run rate outlook: 0.35*** and 0.25***
    - ∆Long-run rate outlook: 0.30*** and 0.25***
    - ∆Log(CDS): 0.052*** and 0.13***
    - ∆US yield: 0.029** and 0.043**
    - ∆ US/PLN 2 year forward rate: -0.80*** and -0.80***
  - Variance equation highlights:
    - GARCH(-1): 0.86*** and 0.87***
    - RESID(-1)^2: 0.19*** and 0.17***
    - ∆Log(VIX) enters variance significantly: 0.0014*** and 0.0021***
  - R-squared examples: 0.373; 0.34

- Table 10: Poland — GARCH including Foreign Ownership (Dependent: Local Currency 5Y Bond Yield)
  - Mean equation:
    - Yield, lag: 0.84*** and 0.87***
    - Foreign ownership: -0.021 (column 1); -0.015*** (column 2)
  - Variance equation:
    - Foreign ownership effect on variance: -0.000041 (column 1); 0.00040*** (column 2)
    - GARCH(-1): 0.65*** and 1.07***
  - R-squared: 0.862 and 0.909

*Source: IMF staff estimates; content unit: _wp1429 - References (Appendixes and Tables as provided).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp1429.pdf_