Abstract
Selected Issues
Indonesia’s Government Bond Yields and Nonresident Participation in Government Bond Markets1
This note presents the evolution of government bond yields in Indonesia and compares it with those in other large Emerging Markets (EMs). The note documents that during periods of global financial stress, nonresident participation in domestic government bond markets in EMs can be fickle while exchange rate volatility can be substantial. The empirical analysis suggests that a larger share of nonresident participation in domestic bond markets seems to be associated with both lower levels of local currency government bond yields and higher sensitivity to changes in global risk aversion. It also finds that both local currency and foreign currency bond yields are positively correlated with exchange rate volatility.
1. This note explores the questions of whether government bond yields in EMs vary with the share of nonresident participation in government bond markets and exchange rate volatility. The first section discusses the evolution in bond yields in Indonesia in recent years and compares it with that of other EMs and global factors. The second section documents the exchange rate volatility in Indonesia and other EMs, especially in episodes of global financial stress. The third section documents the evolution of nonresident participation in domestic bond markets in Indonesia and other EMs, highlighting substantial differences across countries. In the last section, the note presents an econometric analysis of the correlations between government bond yields and bond market characteristics such as the participation by nonresidents in government bond markets, the role of exchange rate volatility, and other domestic and global factors. The objective is to analyze whether the share of nonresident investors in the government bond market is systematically correlated with bond yields, while controlling for domestic factors and time and country specific factors. The latter is an aspect that differentiates this analysis from previous work and contributes to the literature. The note concludes with policy considerations.
2. Indonesian government bond yields have declined in the past decade compared to the early 2000s. However, since the beginning of the 2010s, and in line with the EMs average, Indonesian U.S. dollar bond yields have declined more than the yields on local currency government bonds (Figure 1, left chart). More recently, with the spike in global risk aversion (as measured by the VIX) immediately after the COVID-19 shock hit the global economy, the yields of both Indonesian local currency government and U.S. dollar government bonds increased sharply before declining when external pressures started to ease as well as liquidity was restored in bond markets. However, the yields of Indonesian U.S. dollar government bonds have declined more than yields in local currency government bonds, leading to an increase in the yield differential between these bonds (Figure 1, right chart). These different dynamics in the yields of government bonds depending on currency denomination reflect differences in liquidity in these markets and differences in the perception of currency and credit risk, as well as differences in local currency bond market characteristics, such as the share of nonresident participation in government bond markets.
Government Bond Local Currency and FX Bond Yields
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
3. Global financial conditions tightened after the COVID-19 shock. Government bond yields increased in Indonesia and across all EMs, with market pressures being the sharpest in March 2020 (Figure 2, left chart). As of September 2020, local currency bond yields in Indonesia were at levels comparable to those in other large EMs, such as Brazil, Russia, India, and Mexico, but relatively higher when compared to those in other ASEAN EMs countries, Thailand, Malaysia, and the Philippines (Figure 2, right chart).
Local Currency Bond Yields During the COVID-19 Crisis
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
A. Nonresident Participation in Government Bond Markets
4. The share of nonresident participation in local currency bond markets in Indonesia has increased more than in other EMs since the 2010s (Figure 3, left chart). In 2019, Indonesia was among the EMs with the highest participation of nonresidents in local currency bond markets (Figure 3, right chart). This high, pre-pandemic share also implies that Indonesia represented the second-largest share (just below Mexico) of the total amount invested in EMs local currency bond markets by nonresidents.
Nonresident Participation in Local Currency Government Bond Markets
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
5. A higher participation of nonresidents in government bond markets has benefits but it also entails risks. The increased holdings of EMs government bonds by nonresident investors has contributed to improving government bond markets liquidity in these economies and has also increased financial resources available domestically. In principle, increased participation by nonresidents can lead to lower funding costs as it increases the availability of financial resources in the economy (Peiris, 2010 and Ebeke and Kyobe, 2015). However, a higher share of nonresident holdings of government bonds can also increase the exposure to risks associated with sudden capital outflows (IMF/WB, 2020). Such outflows by nonresidents in periods of financial stress would imply that local currency bond yields would likely be more sensitive to changes in global risk aversion, and potentially translate into disorderly market conditions.
6. Since the beginning of the COVID-19 crisis, the share of nonresident holdings of government bonds has declined across EMs, with relatively large declines in Indonesia. In the first semester of 2020, the share of nonresident holdings in total outstanding debt securities in Indonesia declined from 50 percent to 45 percent, compared to a reduction, on average, from 36 percent to 34 percent in EMs (Figure 4, left chart). The composition by currency denomination in the decline in nonresident participation varied across countries. In Indonesia, Egypt, and Malaysia most of the decline of nonresident participation was accounted for by outflows from local currency bond markets. In Poland and Turkey, on the other hand, the distribution in outflows was even between FX and local currency denominated securities. In Mexico, the decline in the participation of nonresidents in the local currency bond market was the sharpest one among EMs, but it was largely compensated by large issuances of FX debt (Figure 4, right chart). In Indonesia, domestic banks substantially increased their holdings of local currency government bonds, more than compensating for the reduced holdings by nonresident investors.
Nonresident Participation in Total Debt Government Bond Markets During COVID-19 Crisis
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
7. Nonresident participation in EMs local currency bond markets can be particularly fickle during global financial market turmoil. During such periods, the participation of nonresident investors in EMs bond markets has typically declined, leading to disorderly market conditions in local currency bond markets and increases in bond yields. A comparison across different episodes shows that EMs selloffs in the GFC and 2018 were particularly acute for the EMs government bonds asset class (Figure 5, left chart). ASEAN countries also experienced declines in nonresident participation in these episodes, as well as in the 2015 devaluation episode in China (Figure 5, right chart). In the case of Indonesia, the decline in nonresident participation in local currency bond markets in the COVIID-19 crisis exceeded that of other episodes (see text chart on the right).
Nonresident Participation in Local Currency Government Bond Markets During Episodes of Global Financial Stress
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
Indonesia: Change in Share of Nonresident Holdings of LC Government Bonds
(In percent)
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
Sources: Haver Analytics; Bloomberg L.P.; and IMF staff estimates.8. In the COVID-19 crisis, large nonresident outflows from EMs local currency bond markets have been compensated for by sustained issuance of hard-currency bonds by EMs. This asset class has benefited from the ample global liquidity after the monetary easing by the U.S. Fed and other major central banks. Hard-currency bond issuance has been unprecedented in Indonesia, compared to other periods of global financial stress, and they have also outperformed the issuances by other EMs. (Figure 6). In net terms, in 2020 the substitution between nonresident holdings of local currency and foreign currency debt has been substantial. Nonresident outflows from local currency bonds reached about US$9 billion, while bond issuances of Indonesian government bonds denominated in U.S. dollar amounted to about US$10.9 billion (see text chart on the right).
Issuance of Hard Currency Bonds During Episodes of Global Financial Stress
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
Local Currency Bonds Net Outflows and U.S.Dollar Bond Issuances, 2000
(In billions of U.S. dollar)
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
Sources: Haver Analytics; Bloomberg L.P.; and IMF staff estimates.B. Exchange Rate Volatility
9. Sharp exchange rate depreciations tend to lower the take-home returns for nonresident investors, thereby affecting their demand for local currency EMs bonds. Nonresident investors, even when entering the local currency bond markets, are optimizing returns measured in foreign currency (Hofmann, Shim, and Shin, 2020). Changes in the exchange rate, therefore, affect their investment decision, and the tendency toward exchange rate-related, synchronized exit/entry of some nonresident investors can reinforce increases/declines in yields on government bonds denominated in local currency. Also, in periods of global market turbulence, exchange rate volatility, combined with a preference for assets in FX by nonresident investors, can lead to an increase in the yield differential between local currency and FX government bonds. Since the beginning of the COVID-19 crisis, this differential increased in Indonesia, Brazil and Russia but it declined in the Philippines and South Africa, for example (see text chart).
Change in Yield Differential Local Currency vs U.S. Dollar Bonds
(Change December 20l9-August2020, in percentage points)
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
Sources: Bloonnberg L.P.; and IMF staff estimate?.10. In previous episodes of global financial stress, EMs exchange rates depreciated sharply, and exchange rate volatility increased substantially. In relative terms, exchange rate depreciations were the largest in the GFC, but they were also considerable in the COVID-19 crisis (Figure 7). In Indonesia, episodes of increased exchange rate volatility have been mainly associated with global financial stress. Indeed, the volatility of the Rupiah was the largest in the GFC (2008–2009), the Taper Tantrum (2013), the devaluation episode in China (2015), and the COVID-19 crisis (2020) (Figure 8, left chart). Considering the period 2004–2020, exchange rate volatility in Indonesia has been close to the median across EMs, but it has been relatively higher than in other ASEAN peers (Figure 8, right chart).
Exchange Rate Market Pressures
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
Exchange Rate Volatility
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
C. Determinants of Government Bond Yields
11. As noted earlier, government bond yields tend to be associated with country characteristics, such as the participation of nonresidents in government bond markets or the volatility of the exchange rate. In addition, other domestic and global factors also matter. The regression analysis in this note uses government bond yields in local currency and government bond yields in U.S. dollars as dependent variables (Equation 1). This note contributes to the literature by simultaneously estimating the impact of nonresident participation and FX volatility on EMs bond yields while controlling for other domestic and global factors. The econometric model includes the share of nonresident holdings of total government bond securities and the volatility of the lagged daily changes in the exchange rate, and other country-specific variables, such as the monetary policy rate, real GDP growth, inflation, and the returns in the local equity market. The model also includes global factors, such as the VIX and the yields in the 10-year U.S. treasury bonds, as in Miyajima and others (2012), and the returns on U.S. equities. Government bond yields are also likely to be affected by the fiscal stance and debt sustainability concerns. These are not explicitly included in the model due to lack of data availability on a monthly basis. To capture the sensitivity to global factors, given some specific characteristics, the analysis also considers interaction terms (e.g., the share of nonresident holdings of total government debt securities interacted with the VIX). All other country-specific factors that are assumed to be constant over time are captured by country fixed-effects. All other factors varying over time and potentially affecting all countries are captured by the inclusion of time fixed effects. The bond yields, the monetary policy rate, and the global factors enter the equation in logs. The regressions use quarterly data over the period 2004:Q1–2020:Q2. The countries included in the analysis correspond to a sample of 18 large EMs, selected based on data availability, most notably on high-frequency information on nonresident participation in government bond markets from Arslanalp and Tsuda (2014) (see Appendix 1).
Bond yieldsit = Nonresidentit-1 + Volatility (Daily FX Changesit-1) + Other Domestic Factorsit-1 + VIXit-1+ US 10yr yieldit-1 + Interactions terms + time FE + country FE (1)
12. The regression analysis corroborates that local currency government bond yields are affected by domestic and global factors. The econometric analysis shows that, on average, local currency bond yields are negatively associated with a higher presence of nonresidents in bond markets (Table 1). This result is consistent with the view that the increased availability of resources lowers funding costs. However, positive coefficient on the interaction term between the share of nonresident holdings of total government bond securities and the VIX also shows that this benefit comes at the cost of a higher exposure to global risk perceptions. If the latter, as captured by the VIX, increase, a higher share of nonresident participation in government bonds is associated with higher local currency yields in EMs (Figure 9, chart). This finding highlights the tradeoffs associated with a higher share of nonresident participation. More specifically, the regression analysis shows that, at the average sample level of the VIX, a one-standard deviation increase in the share of nonresident participation in government bonds is associated with an increase of 1.0 percent in the average level of local currency bond yield. The regression results also suggest that an increase in exchange rate volatility is associated with higher local currency bond yields in EMs. An increase of one-standard deviation in the exchange rate volatility of the exchange rate is associated with an increase of 1.03 percentage points in local currency bond yields. This positive correlation is consistent with the hypothesis that nonresident investors require higher returns to compensate for higher exchange rate volatility. The econometric results also show that local currency government bond yields are positively associated with monetary policy rates. The coefficients are robust to different specifications, including sequentially excluding one of the main variables of interest.
Emerging Markets: Determinants of Local Currency Government Bond Yields 1/
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
Emerging Markets: Determinants of Local Currency Government Bond Yields 1/
| Local Currency Bond Yields | ||
|---|---|---|
| VARIABLES | (1) | (2) |
| Nonresident in percent of total bonds (t-1) | -0.009 *** (0.003) |
-0.009 *** (0.003) |
| Nonresident in percent of total bonds (t-1) x Vix (t-1) | 0.004 *** (0.001) |
0.004 *** (0.001) |
| FX daily volatility (t-1) | 0.063 ** (0.030) |
0.063 ** (0.030) |
| Monetary policy rate (t-1) | 0.334 *** (0.040) |
0.334 *** (0.040) |
| Real GDP growth (t-1) | -0.610 (0.427) |
-0.610 (0.427) |
| Inflation (t-1) | 0.067 (0.652) |
0.067 (0.652) |
| Equity index, percent change (t-1) | -0.034 (0.049) |
-0.034 (0.049) |
| VIX (t-1) | -0.203 (0.117) |
-0.187 * (0.103) |
| US treasury 10-year bond yield (t-1) | -0.088 (0.159) |
-0.047 (0.118) |
| US equity index, percent change (t-1) | -0.195 (0.291) |
|
| Constant | 1.594 *** (0.371) |
1.557 *** (0.339) |
| Observations | 1,023 | 1,023 |
| Time fixed-effects | Yes | Yes |
| Country fixed-effects | Yes | Yes |
| R-squared | 0.692 | 0.692 |
| Number of c_code | 18 | 18 |
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
Emerging Markets: Determinants of Local Currency Government Bond Yields 1/
| Local Currency Bond Yields | ||
|---|---|---|
| VARIABLES | (1) | (2) |
| Nonresident in percent of total bonds (t-1) | -0.009 *** (0.003) |
-0.009 *** (0.003) |
| Nonresident in percent of total bonds (t-1) x Vix (t-1) | 0.004 *** (0.001) |
0.004 *** (0.001) |
| FX daily volatility (t-1) | 0.063 ** (0.030) |
0.063 ** (0.030) |
| Monetary policy rate (t-1) | 0.334 *** (0.040) |
0.334 *** (0.040) |
| Real GDP growth (t-1) | -0.610 (0.427) |
-0.610 (0.427) |
| Inflation (t-1) | 0.067 (0.652) |
0.067 (0.652) |
| Equity index, percent change (t-1) | -0.034 (0.049) |
-0.034 (0.049) |
| VIX (t-1) | -0.203 (0.117) |
-0.187 * (0.103) |
| US treasury 10-year bond yield (t-1) | -0.088 (0.159) |
-0.047 (0.118) |
| US equity index, percent change (t-1) | -0.195 (0.291) |
|
| Constant | 1.594 *** (0.371) |
1.557 *** (0.339) |
| Observations | 1,023 | 1,023 |
| Time fixed-effects | Yes | Yes |
| Country fixed-effects | Yes | Yes |
| R-squared | 0.692 | 0.692 |
| Number of c_code | 18 | 18 |
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
Impact of Nonresident Participation and FX Volatility on Local Currency Government Bond Yields
Citation: IMF Staff Country Reports 2021, 047; 10.5089/9781513570860.002.A001
13. The regressions also suggest that yields of government bonds denominated is U.S. dollar are associated with different factors than the bond yields in local currency bonds. Bond yields in foreign currency are positively correlated with global factors such as the VIX and the returns of U.S. treasury bonds, and do not respond to the monetary policy rate or the nonresident participation in government bond markets (as it is the case of local currency bond yields) (Table 2). This positive correlation is in line with the intuition that as global risk aversion increases as measured by the VIX investors, demand higher yields for their (risky) investments. Also, when the yields of U.S. treasury bonds increase, then yields for risky assets denominated in U.S. dollars also increase as they are perceived as alternative investments with lower credit quality. On the other hand, bond yields in U.S. dollar denominated bonds decline in the presence of strong economic activity, consistent with the view that investor appetite for bonds is positively correlated with strong economic performance. In addition, government bond yields in U.S. dollar are positively associated with the volatility of the exchange rate, as is the case of local currency government bond yields.
Emerging Markets: Determinants of U.S. Dollar Government Bond Yields 1/
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
Emerging Markets: Determinants of U.S. Dollar Government Bond Yields 1/
| U.S. Dollar Bond Yields | ||
|---|---|---|
| VARIABLES | (1) | (2) |
| Nonresident in percent of total bonds (t-1) | -0.004 (0.005) |
-0.004 (0.005) |
| Nonresident in percent of total bonds (t-1) x Vix (t-1) | 0.003 (0.002) |
0.003 (0.002) |
| FX daily volatility (t-1) | 0.141 *** (0.026) |
0.141 *** (0.026) |
| Monetary policy rate (t-1) | 0.076 (0.055) |
0.076 (0.055) |
| Real GDP growth (t-1) | -1.279 *** (0.421) |
-1.279 *** (0.421) |
| Inflation (t-1) | 0.359 (0.249) |
0.359 (0.249) |
| Equity index, percent change (t-1) | 0.002 (0.061) |
0.002 (0.061) |
| VIX (t-1) | 0.321 ** (0.143) |
0.339 ** (0.128) |
| US treasury 10-year bond yield (t-1) | 0.656 *** (0.158) |
0.702 *** (0.159) |
| US equity index, percent change (t-1) | -0.219 (0.304) |
|
| Constant | -0.235 (0.436) |
-0.277 (0.409) |
| Observations | 1,036 | 1,036 |
| Time fixed-effects | Yes | Yes |
| Country fixed-effects | Yes | Yes |
| R-squared | 0.641 | 0.641 |
| Number of c_code | 18 | 18 |
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
Emerging Markets: Determinants of U.S. Dollar Government Bond Yields 1/
| U.S. Dollar Bond Yields | ||
|---|---|---|
| VARIABLES | (1) | (2) |
| Nonresident in percent of total bonds (t-1) | -0.004 (0.005) |
-0.004 (0.005) |
| Nonresident in percent of total bonds (t-1) x Vix (t-1) | 0.003 (0.002) |
0.003 (0.002) |
| FX daily volatility (t-1) | 0.141 *** (0.026) |
0.141 *** (0.026) |
| Monetary policy rate (t-1) | 0.076 (0.055) |
0.076 (0.055) |
| Real GDP growth (t-1) | -1.279 *** (0.421) |
-1.279 *** (0.421) |
| Inflation (t-1) | 0.359 (0.249) |
0.359 (0.249) |
| Equity index, percent change (t-1) | 0.002 (0.061) |
0.002 (0.061) |
| VIX (t-1) | 0.321 ** (0.143) |
0.339 ** (0.128) |
| US treasury 10-year bond yield (t-1) | 0.656 *** (0.158) |
0.702 *** (0.159) |
| US equity index, percent change (t-1) | -0.219 (0.304) |
|
| Constant | -0.235 (0.436) |
-0.277 (0.409) |
| Observations | 1,036 | 1,036 |
| Time fixed-effects | Yes | Yes |
| Country fixed-effects | Yes | Yes |
| R-squared | 0.641 | 0.641 |
| Number of c_code | 18 | 18 |
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
14. The spread between local currency bond yields and FX bond yields depends crucially on the domestic monetary policy rate, the inflation rate, and the volatility of the exchange rate. We proxy the currency risk premium by the yield differential between the government bond yield in local currency and its correspondent yield in U.S. dollar. We find that this “currency spread” is positively correlated with the monetary policy rate and the inflation rate (Table 3, columns 1–2). The positive correlation with the monetary policy rate can be explained by the fact that the monetary policy rates often operates as the floor for interest rates in the economy. The result on inflation is consistent with the notion that investors are more concerned about real (net on inflation) than nominal returns on their investment. In addition, the econometric results show that the “currency spread” is positively correlated with the volatility of the exchange rate. As mentioned above, this result is in line with the idea that nonresident investors demand higher yields in the presence of uncertain returns measured in their home currency.
Emerging Markets: Determinants of Country Credit Risk and Currency Risk 1/
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
Emerging Markets: Determinants of Country Credit Risk and Currency Risk 1/
| Yield Differential Between LC and FX | EMBI Spread | |||
|---|---|---|---|---|
| VARIABLES | (1) | (2) | (3) | (4) |
| Nonresident in percent of total bonds (t-1) | -0.019 (0.030) |
-0.019 (0.030) |
-0.019 (0.013) |
-0.019 (0.013) |
| Nonresident in percent of total bonds (t-1) x Vix (t-1) | 0.007 (0.011) |
0.007 (0.011) |
0.01 ** (0.005) |
0.01 ** (0.005) |
| FX daily volatility (t-1) | 0.259 ** (0.120) |
0.259 ** (0.120) |
0.282 *** (0.093) |
0.282 *** (0.093) |
| Monetary policy rate (t-1) | 0.518 *** (0.157) |
0.518 *** (0.157) |
0.175 * (0.091) |
0.175 * (0.091) |
| Real GDP growth (t-1) | 2.978 (3.608) |
2.978 (3.608) |
-3.733 ** (1.389) |
-3.733 ** (1.389) |
| Inflation (t-1) | 6.498 *** (1.923) |
6.498 *** (1.923) |
1.290 (1.496) |
1.290 (1.496) |
| Equity index, percent change (t-1) | -0.162 (0.248) |
-0.162 (0.248) |
0.082 (0.159) |
0.082 (0.159) |
| VIX (t-1) | -0.412 (0.920) |
-0.333 (0.918) |
1.008 *** (0.216) |
0.845 *** (0.195) |
| US treasury 10-year bond yield (t-1) | -0.529 (1.185) |
-0.324 (1.163) |
1.376 *** (0.393) |
0.952 ** (0.376) |
| US equity index, percent change (t-1) | -0.970 (1.051) |
2.004 ** (0.706) |
||
| Constant | 0.509 (2.900) |
0.326 (2.894) |
1.058 (0.783) |
1.435 * (0.718) |
| Observations | 759 | 759 | 759 | 759 |
| Time fixed-effects | Yes | Yes | Yes | Yes |
| Country fixed-effects | Yes | Yes | Yes | Yes |
| R-squared | 0.279 | 0.279 | 0.492 | 0.492 |
| Number of c_code | 18 | 18 | 18 | 18 |
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
Emerging Markets: Determinants of Country Credit Risk and Currency Risk 1/
| Yield Differential Between LC and FX | EMBI Spread | |||
|---|---|---|---|---|
| VARIABLES | (1) | (2) | (3) | (4) |
| Nonresident in percent of total bonds (t-1) | -0.019 (0.030) |
-0.019 (0.030) |
-0.019 (0.013) |
-0.019 (0.013) |
| Nonresident in percent of total bonds (t-1) x Vix (t-1) | 0.007 (0.011) |
0.007 (0.011) |
0.01 ** (0.005) |
0.01 ** (0.005) |
| FX daily volatility (t-1) | 0.259 ** (0.120) |
0.259 ** (0.120) |
0.282 *** (0.093) |
0.282 *** (0.093) |
| Monetary policy rate (t-1) | 0.518 *** (0.157) |
0.518 *** (0.157) |
0.175 * (0.091) |
0.175 * (0.091) |
| Real GDP growth (t-1) | 2.978 (3.608) |
2.978 (3.608) |
-3.733 ** (1.389) |
-3.733 ** (1.389) |
| Inflation (t-1) | 6.498 *** (1.923) |
6.498 *** (1.923) |
1.290 (1.496) |
1.290 (1.496) |
| Equity index, percent change (t-1) | -0.162 (0.248) |
-0.162 (0.248) |
0.082 (0.159) |
0.082 (0.159) |
| VIX (t-1) | -0.412 (0.920) |
-0.333 (0.918) |
1.008 *** (0.216) |
0.845 *** (0.195) |
| US treasury 10-year bond yield (t-1) | -0.529 (1.185) |
-0.324 (1.163) |
1.376 *** (0.393) |
0.952 ** (0.376) |
| US equity index, percent change (t-1) | -0.970 (1.051) |
2.004 ** (0.706) |
||
| Constant | 0.509 (2.900) |
0.326 (2.894) |
1.058 (0.783) |
1.435 * (0.718) |
| Observations | 759 | 759 | 759 | 759 |
| Time fixed-effects | Yes | Yes | Yes | Yes |
| Country fixed-effects | Yes | Yes | Yes | Yes |
| R-squared | 0.279 | 0.279 | 0.492 | 0.492 |
| Number of c_code | 18 | 18 | 18 | 18 |
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.
15. The econometric analysis suggests that the proxy for country risk premium is more closely correlated with global factors than the proxy for currency risk premium. We proxy the country risk premium by the EMBI spread, which is equal to the difference between the yield of country i government bond in U.S. dollars relative to yield of the U.S. government bond. The regressions show that, as opposed to the currency risk premium, this premium is positively correlated with global factors, such as risk aversion (VIX), and with alternative investments is U.S. dollars such as U.S. government bonds and the U.S. equity index (Table 3, columns 3–4). This result is in line with the intuition that that the “country risk premium” is compensating for differences in the riskiness of assets denominated in the same currency, the U.S. dollars in our case.
16. A high participation of nonresidents in domestic bond markets has benefits, such as increased access to financing, but it also entails risks as the recent COVID-19 crisis has shown. A higher participation by nonresidents in EMs government bond markets provide for capital flows that can have substantial benefits for countries, including by enhancing efficiency, promoting financial sector competitiveness, and facilitating greater productive investment and consumption smoothing (IMF, 2012). However, the flipside is that higher nonresident participation also increases the exposure to the downside risks of capital outflows in times of stress. In such episodes, capital outflows can amplify rather than counter domestic disorderly market conditions. These vulnerabilities associated with domestic government bond markets could be reduced by strengthening domestic markets infrastructure, including hedging markets and liquidity, broadening inclusion in global bond indices, and maintaining a stable macroeconomic environment (IMF/WB, 2020). The results presented in this note also suggest that keeping inflation low can contribute to a lower currency risk premium, consistent with the important role of sound policies in managing macroeconomic and financial stability risks (IMF, 2012). In addition, in the case of Indonesia, increasing revenue mobilization, deepening domestic bank and nonbank financial markets, e.g., by increasing saving rates and further developing pension funds’ investments in government bonds, and attracting more FDI would contribute to offering alternative sources of financing for the government and reducing the reliance on volatile nonresident portfolio flows into government bonds.
Appendix I. Data Sources
Emerging Markets: Variables and Data Sources
Emerging Markets: Variables and Data Sources
| Variable | Description | Source |
|---|---|---|
| Government 10-year local currency bond yields | In percent | Bloomberg LP. |
| Government 10-year EMBI bond yields | In percent | Bloomberg LP. |
| Nonresident participation in government bond markets | Nonresident holdings of government bands in percent of total outstanding | Bloomberg LP. and author’s calculations |
| FX volatility | Sandard deviation of the lagged daily changes over each quarter | IMF, International Financial Statistics (author’s calculations) |
| Monetary Policy rates | In percent | Haver Analytics |
| Real GDP growth | In percent | IMF, International Financial Statistics |
| Inflation | In percent | Haver Analytics |
| Equity index, percentage change | In percent | Bloomberg LP. |
| VIX | Chicago Board Options Exchange’s CBOE Volatility Index | Bloomberg LP. |
| U.S. 10-year treasury bonds yields | In percent | Bloomberg LP. |
| S&P 500 equity index, percentage change | In percent | Bloomberg LP. |
Emerging Markets: Variables and Data Sources
| Variable | Description | Source |
|---|---|---|
| Government 10-year local currency bond yields | In percent | Bloomberg LP. |
| Government 10-year EMBI bond yields | In percent | Bloomberg LP. |
| Nonresident participation in government bond markets | Nonresident holdings of government bands in percent of total outstanding | Bloomberg LP. and author’s calculations |
| FX volatility | Sandard deviation of the lagged daily changes over each quarter | IMF, International Financial Statistics (author’s calculations) |
| Monetary Policy rates | In percent | Haver Analytics |
| Real GDP growth | In percent | IMF, International Financial Statistics |
| Inflation | In percent | Haver Analytics |
| Equity index, percentage change | In percent | Bloomberg LP. |
| VIX | Chicago Board Options Exchange’s CBOE Volatility Index | Bloomberg LP. |
| U.S. 10-year treasury bonds yields | In percent | Bloomberg LP. |
| S&P 500 equity index, percentage change | In percent | Bloomberg LP. |
Emerging Markets: Sample of Countries in the Regression Analysis
Emerging Markets: Sample of Countries in the Regression Analysis
| List of Countries | |||
|---|---|---|---|
| Brazil | Chile | China | Colombia |
| Hungary | India | Indonesia | Malaysia |
| Mexico | Peru | Philippines | Poland |
| Romania | Russia | South Africa | Thailand |
| Turkey | Ukraine | ||
Emerging Markets: Sample of Countries in the Regression Analysis
| List of Countries | |||
|---|---|---|---|
| Brazil | Chile | China | Colombia |
| Hungary | India | Indonesia | Malaysia |
| Mexico | Peru | Philippines | Poland |
| Romania | Russia | South Africa | Thailand |
| Turkey | Ukraine | ||
References
Arslanalp, S. and T. Tsuda, 2014, “Tracking Global Demand for Emerging Market Sovereign Debt,” IMF Working Paper No. 14/39 (Washington: International Monetary Fund).
Ebeke, C. and A. Kyobe, 2015, “Global Financial Spillovers to Emerging Market Sovereign Bond Markets,” IMF Working Paper No. 15/141 (Washington: International Monetary Fund).
Hofmann, B., I. Shim, and H.S. Shin, 2020, “Bond Risk Premia and the Exchange Rate,” BIS Working Paper No. 775 (Basel: Bank for International Settlements).
International Monetary Fund and World Bank, 2020, “Recent Development in Local Currency Bond Markets in Emerging Economies,” in Staff Note for the G-20 International Financial Architecture Working Group, January (Riyadh, Kingdom of Saudi Arabia).
International Monetary Fund, 2012, “The Liberalization and Management of Capital Flows: An Institutional View,” IMF Policy Paper. Available via the Internet: https://www.imf.org/-/media/Websites/IMF/imported-full-text-pdf/external/np/pp/eng/2012/_111412.ashx.
Miyajima, K, M. S. Mohanty, and T. Chan, 2012, “Emerging Market Local Currency Bonds: Diversification and Stability,” BIS Working Paper No. 391 (Basel: Bank for International Settlements).
Peiris, S., 2010, “Foreign Participation in Emerging Markets’ Local Currency Bond Markets,” IMF Working Paper No. 10/88 (Washington: International Monetary Fund).
Prepared by Francisco Arizala (SPR).
