## 1. MSBFs Assets under Management

## Source details

**Canonical URL:** [1. MSBFs Assets under Management](https://www.imf.org/-/media/files/publications/gfs-notes/2021/english/gfsnea2021005.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/gfs-notes/2021/english/gfsnea2021005.pdf.md)
- [Structured JSON version](/-/media/files/publications/gfs-notes/2021/english/gfsnea2021005.pdf.json)

---

### Concentrated emerging market exposures
- Sample: 40 large MSBFs, domiciled in the United States and Europe, from the fourth quarter of 2009 to the second quarter of 2020.
- Over the past decade the sample of 40 MSBFs accounted for more than 20 percent of the foreign investor base of sovereign bond markets in countries such as Hungary, Malaysia, Ukraine, and Uruguay; more recently they exceeded 10 percent of the foreign investor base in Brazil, India, and Mexico.
- MSBFs reached historical lows in investment concentration in many emerging markets as of the second quarter of 2020, having rebalanced into safer assets and sold positions in response to redemptions.
- Local bond markets and currencies of some affected emerging markets underperformed broader emerging market indices during this period, with Brazil and India suffering particularly large price pressure.

### Benchmark diversion, activism, and risk profile
- MSBFs exhibit a high degree of activism: median active share exceeding 70 percent during “normal” times (fourth quarter of 2009 to second quarter of 2018, per Cortes and Sanfilippo 2020).
- By contrast, benchmark-driven investors (BDIs) and emerging market economy funds proxied by EPFR global bond funds and EPFR-dedicated emerging markets funds have active shares below 20 percent.
- Active share methodology: Cremers and Petajisto (2009) comparing fund holdings to J.P. Morgan emerging markets debt indices (EMBIG and GBI-EM Global Diversified).
- MSBFs’ unconstrained mandates can permit holding positions through volatility shocks, potentially stabilizing; however, observed behavior can be opportunistic and unstable due to frequent rotations.

### Changes in active share and currency breakdown (2018–2020)
- Overall weighted active share for MSBFs fell since 2018, with the drop accelerating after the COVID-19 shock.
- The drop in active share was larger for hard currency than local currency: hard currency active share fell by 3.8 percent versus 1.3 percent for local currency in 2020.
- MSBFs have had nearly 70 percent of their emerging market investment in local currency over the past decade.

### Portfolio allocation changes during COVID-19
- MSBFs experienced portfolio reallocation estimated as outflows of $23 billion during the COVID-19 outbreak, almost entirely out of local currency bond exposures.
- Reallocations were concentrated in local currency bonds of large, liquid emerging market sovereign issuers (Brazil and India); exposures to less liquid emerging market economies (such as Argentina) remained largely unchanged.
- For Latin America and the Caribbean (LAC) region: divestments were concentrated in Brazilian local currency bonds; hard currency exposures remained largely unchanged.

### Redemption analysis and dynamics during stress
- Redemptions defined as returns adjusted for changes in assets under management (AUM).
- MSBFs, as a group, suffered large redemptions during the first quarter of 2020—larger than in any other period in the sample (fourth quarter of 2009 to second quarter of 2020).
- Periods of MSBF redemptions (isolated, clustered, or system-wide) coincided with overall risk aversion, suggesting forced liquidations that tend to exacerbate cross-border flows rather than stabilize them.

### Regression analysis: impact of MSBF reallocations
- Empirical framework: three-dimensional panel regressions with fund-country fixed effects; main explanatory variable is the first difference log of MSBF emerging market fixed income holdings (price-adjusted) with interaction by a risk appetite dummy (DRA: 0 = risk-off, 1 = risk-on).
- Dependent variables (performance spreads, expressed in percent; positive = domestic asset outperforms benchmark):
  - Currency Spread (CrncySP) = quarter-over-quarter change in local currency vs. US dollar minus quarter-over-quarter change in the J.P. Morgan Emerging Market Currency Index.
  - Local Currency Government Bond Index Spread (GBISP) = quarter-over-quarter change in domestic Local Currency Total Return Bond index minus quarter-over-quarter change in the J.P. Morgan Government Bond Index-Emerging Markets.
  - Hard Currency Government Bond Index Spread (EMBIGSP) = quarter-over-quarter change in domestic Hard Currency Total Return Bond index minus quarter-over-quarter change in the J.P. Morgan Emerging Market Bond Index.
- Key regression results (Estimation period: quarterly data from the fourth quarter of 2009 to the second quarter of 2020):
  - Panel A. CrncySP (Obs. 4,668; R2 0.205)
    - dlog(msbf): 1.65      ***
      - Standard error: (0.30)
    - DRA: 0.06
      - Standard error: (0.42)
    - dlog(msbf) x DRA: -1.35 ***
      - Standard error: (0.44)
  - Panel B. GBISP (Obs. 4,339; R2 0.163)
    - dlog(msbf): 1.87      ***
      - Standard error: (0.39)
    - DRA: -0.76
      - Standard error: (0.57)
    - dlog(msbf) x DRA: -1.41 **
      - Standard error: (0.55)
  - Panel C. EMBIGSP (Obs. 19,769; R2 0.049)
    - dlog(msbf): -0.78      ***
      - Standard error: (0.26)
    - DRA: 0.14
      - Standard error: (0.13)
    - dlog(msbf) x DRA: 0.73 **
      - Standard error: (0.30)
  - Significance legend: ***, **, * indicate significance at the 1, 5, and 10 percent levels, respectively.

### Estimated impacts and interpretation
- Currencies and local currency bonds:
  - Positive and statistically significant relation between changes in MSBF holdings and currency and local currency bond performance.
  - In risk-off periods (DRA = 0), an increase of MSBF outflows in an emerging country allocation of 1 percent is associated with underperformance of:
    - local currency vs EMCI by –1.7 percent, and
    - local currency bonds vs GBI benchmark by –1.9 percent.
  - In periods of inflows, these impacts are moderated to –0.3 percent and –0.5 percent, respectively.
- Hard currency bonds:
  - Coefficients significant but with a counterintuitive sign: an increase of MSBF outflows in an emerging country allocation of 1 percent is associated with outperformance over the EMBIG benchmark by +0.8 percent; in a period of inflows this impact is moderated to +0.1 percent.
  - Interpretation: MSBFs’ hard currency holdings were more resilient during COVID-19 and recovered rapidly, complicating isolation of MSBF flow effects in empirical estimation.

### Advanced economy versus emerging market fixed income dynamics
- MSBF advanced economy exposures are larger overall, but emerging market portfolios bore the brunt during stress periods.
- Quarter-over-quarter percent changes (price-adjusted) show significantly larger emerging market outflows relative to advanced economy portfolios in the second half of 2018 and the first quarter of 2020—especially for local currency versus hard currency.
- When advanced economy markets calmed, MSBFs continued pulling out of emerging market local currency while recovering some emerging market hard currency exposures.
- Conclusion: MSBFs can retreat disproportionately from selected emerging market exposures; during large redemptions they tend to sell their largest emerging market positions (often local currency), which can amplify instability rather than stabilize emerging market fixed income markets.

### Economic interpretation and linkage to the literature
- Findings align with literature highlighting "original sin redux": borrowing in local currency mitigates borrower currency mismatch but transfers mismatch to lenders, exposing foreign investors to currency risk (Carstens and Shin 2019; BIS 2019).
- During risk aversion episodes, weakening local currencies and domestic asset sell-offs prompt foreign investors to reduce exposure, triggering outflows.
- Increased foreign currency hedging can exert additional pressure on exchange rates and funding costs even absent outright outflows.
- Emerging market exchange rates can amplify portfolio adjustments due to pro-cyclicality: local currency spreads and exchange rates move in lockstep.
- Conclusion: local currency issuance aids sovereigns in developing domestic markets but entails risks when a large share of investors are opportunistic and concentrated foreign investors.

*Source: gfsnea2021005 - 1. MSBFs Assets under Management*

### 1. MSBFs Assets under Management

### 1. MSBFs Assets under Management

### Concentrated Emerging Market Exposures
- Sample: 40 large MSBFs, domiciled in the United States and Europe, from the fourth quarter of 2009 to the second quarter of 2020.
- Over the past decade the sample of 40 MSBFs accounted for more than 20 percent of the foreign investor base of sovereign bond markets in countries such as Hungary, Malaysia, Ukraine, and Uruguay; more recently they exceeded 10 percent of the foreign investor base in Brazil, India, and Mexico.
- MSBFs reached historical lows in investment concentration in many emerging markets as of the second quarter of 2020, having rebalanced into safer assets and sold positions in response to redemptions.
- Local bond markets and currencies of some affected emerging markets underperformed broader emerging market indices during this period, with Brazil and India suffering particularly large price pressure.

### Benchmark Diversion and Risks
- MSBFs exhibit a high degree of activism: median active share exceeding 70 percent during “normal” times (fourth quarter of 2009 to second quarter of 2018, per Cortes and Sanfilippo 2020).
- By contrast, benchmark-driven investors (BDIs) and emerging market economy funds proxied by EPFR global bond funds and EPFR-dedicated emerging markets funds have active shares below 20 percent.
- Active share methodology: Cremers and Petajisto (2009) comparing fund holdings to J.P. Morgan emerging markets debt indices (EMBIG and GBI-EM Global Diversified).
- MSBFs’ unconstrained mandates can permit holding positions through volatility shocks, potentially stabilizing; however, their observed behavior can also be opportunistic and unstable due to frequent rotations.

### Changes in Active Share and Currency Breakdown (2018–2020)
- Overall, MSBF weighted active share fell since 2018, with the drop accelerating after the COVID-19 shock.
- The drop in active share was larger for hard currency than local currency: hard currency active share fell by 3.8 percent versus 1.3 percent for local currency in 2020.
- MSBFs have had nearly 70 percent of their emerging market investment in local currency over the past decade.

### Portfolio Allocation Changes During COVID-19
- MSBFs experienced portfolio reallocation estimated as outflows of $23 billion during the COVID-19 outbreak, almost entirely out of local currency bond exposures.
- Reallocations were concentrated in local currency bonds of large, liquid emerging market sovereign issuers (Brazil and India), while exposures to less liquid emerging market economies (such as Argentina) remained largely unchanged.
- For Latin America and the Caribbean (LAC) region: divestments were concentrated in Brazilian local currency bonds; hard currency exposures remained largely unchanged.

### Redemption Analysis
- Redemptions defined as returns adjusted for changes in assets under management (AUM).
- MSBFs, as a group, suffered large redemptions during the first quarter of 2020—larger than in any other period in the sample (fourth quarter of 2009 to second quarter of 2020).
- Periods of MSBF redemptions (isolated, clustered, or system-wide) coincided with overall risk aversion, suggesting forced liquidations that tend to exacerbate cross-border flows rather than stabilize them.

### Regression Analysis: Impact of MSBF Reallocations
- Empirical approach: panel regressions where dependent variables are performance spreads over emerging market benchmarks:
  - Currency: performance spread of local currency over the J.P. Morgan Emerging Market Currency Index (EMCI).
  - Local currency bonds: performance relative to the J.P. Morgan GBI Diversified Index.
  - Hard currency bonds: performance relative to the J.P. Morgan EMBI-Global Diversified Index.
- Main explanatory variable: first difference log of MSBF emerging market fixed income holdings, adjusted for valuation changes; interactions include a dummy for risk appetite (0 = risk-off, 1 = risk-on).
- Key regression results:
  - Currencies and local currency bonds: positive relation between changes in MSBF holdings and performance; in risk-off periods, an increase of MSBF outflows in an emerging country allocation of 1 percent is associated with underperformance of:
    - local currency vs EMCI by –1.7 percent, and
    - local currency bonds vs GBI benchmark by –1.9 percent.
    - In periods of inflows, these impacts are moderated to –0.3 percent and –0.5 percent, respectively.
  - Hard currency bonds: coefficients significant but with a counterintuitive sign: an increase of MSBF outflows in an emerging country allocation of 1 percent is associated with outperformance over the EMBIG benchmark by +0.8 percent; in a period of inflows this impact is moderated to +0.1 percent.
- Interpretation for hard currency result: MSBFs’ hard currency holdings were more resilient during COVID-19 and recovered rapidly, complicating isolation of MSBF flow effects in empirical estimation.

### Advanced Economy vs Emerging Market Fixed Income Dynamics
- MSBF advanced economy exposures are larger overall, but emerging market portfolios bore the brunt during stress periods.
- Table 1 quarter-over-quarter percent changes (price-adjusted) show significantly larger emerging market outflows relative to advanced economy portfolios in the second half of 2018 and the first quarter of 2020—especially for local currency versus hard currency.
- When advanced economy markets calmed, MSBFs continued pulling out of emerging market local currency while recovering some emerging market hard currency exposures.
- Conclusion: MSBFs can retreat disproportionately from selected emerging market exposures; during large redemptions they tend to sell their largest emerging market positions (often local currency), which can amplify instability rather than stabilize emerging market fixed income markets.

*Source: gfsnea2021005 - 1. MSBFs Assets under Management*

### References

### gfsnea2021005 - References

### Regression analysis: methodology and sample
- Sample extended through the second quarter of 2020 to include the COVID-19 shock and ensuing overall emerging market flows bust and nascent recovery in hard currency.
- Builds on the panel regression analysis first presented in Cortes and Sanfilippo (2020) using the same parsimonious three-dimensional panel regressions with interaction effects.
- Cross-section identifiers:
  - (1) MSBF (multi-sector bond fund)
  - (2) country invested in
  - Third dimension: time (quarterly data).
- Main regression specification:
  - yijt = xijt′β + αij + uijt
  - αij are fund-country fixed effects; uijt are time-variant unobservables.
- Estimation issues and corrections:
  - Estimations account for cross-sectional dependence and heteroscedasticity.
  - An F-test supports panel fixed effects over pooled OLS; time fixed effects were tested as robustness but broadly insignificant and not included in presented results.
  - Standard errors and covariances estimated using a degree-of-freedom corrected variant of the Panel Corrected Standard Error (PCSE) methodology (Beck and Katz 1995), employing a SUR-like handling of cross-section correlation.

### Variables and definitions
- Dependent variables (performance spreads, expressed in percent; positive = domestic asset outperforms benchmark):
  - Currency Spread (CrncySP) = (Quarter-over-quarter change in local currency vs. US dollar) – (Quarter-over-quarter change in the J.P. Morgan Emerging Market Currency Index).
  - Local Currency Government Bond Index Spread (GBISP) = (Quarter-over-quarter change in domestic Local Currency Total Return Bond index) – (Quarter-over-quarter change in the total return overall on the J.P. Morgan Government Bond Index-Emerging Markets).
  - Hard Currency Government Bond Index Spread (EMBIGSP) = (Quarter-over-quarter change in domestic Hard Currency Total Return Bond index) – (Quarter-over-quarter change in the total return overall on the J.P. Morgan Emerging Market Bond Index).
- Explanatory variables:
  - MSBFs Flows (dlog(msbf)): first difference log of price-adjusted dollar market value of MSBF portfolio allocations per fund per country (local currency holdings used for panels A and B; hard currency holdings used for panel C). Price-adjustment procedure described in four steps preserving index return adjustments.
  - Risk Appetite dummy (DRA): equals 1 in quarters of net emerging market portfolio inflows (risk-on), and 0 in periods of outflows (risk-off). Interaction term dlog(msbf) x DRA included to test differential impacts in risk-on versus risk-off periods.

### Key estimation results and statistical details (Table 2)
- Estimation period: quarterly data from the fourth quarter of 2009 to the second quarter of 2020. Estimations include fund-country fixed effects.
- Panel A. CrncySP (Obs. 4,668; R2 0.205)
  - dlog(msbf): 1.65      ***
    - Standard error: (0.30)
  - DRA: 0.06
    - Standard error: (0.42)
  - dlog(msbf) x DRA: -1.35 ***
    - Standard error: (0.44)
- Panel B. GBISP (Obs. 4,339; R2 0.163)
  - dlog(msbf): 1.87      ***
    - Standard error: (0.39)
  - DRA: -0.76
    - Standard error: (0.57)
  - dlog(msbf) x DRA: -1.41 **
    - Standard error: (0.55)
- Panel C. EMBIGSP (Obs. 19,769; R2 0.049)
  - dlog(msbf): -0.78      ***
    - Standard error: (0.26)
  - DRA: 0.14
    - Standard error: (0.13)
  - dlog(msbf) x DRA: 0.73 **
    - Standard error: (0.30)
- Significance legend: ***, **, * indicate significance at the 1, 5, and 10 percent levels, respectively.

### Interpretation of results (extracted conclusions)
- Currency and local currency fixed income regressions:
  - Positive and statistically significant relation between changes in MSBF holdings and (i) currency and (ii) local currency bond performance.
  - In risk-off periods (DRA = 0), an increase of MSBF inflows by 1 percent in an emerging market is associated with:
    - (i) currency outperformance over the EMCI by 1.65 percent (vs. 1.09 percent in C&S).
    - (ii) local currency bond outperformance over the GBI benchmark by 1.87 percent (vs. 1.23 percent in C&S).
  - In risk-on periods (DRA = 1), the impact is moderated to:
    - (i) 0.30 percent (vs. 0.44 percent in C&S).
    - (ii) 0.47 percent (vs. 0.41 in C&S).
- Hard currency fixed income regression:
  - Coefficients significant but sign is negative for dlog(msbf).
  - In risk-off periods (DRA = 0), an increase of MSBF inflows by 1 percent is associated with underperformance versus the EMBIG benchmark by –0.78 percent (vs. +0.17 percent in C&S).
  - In risk-on periods (DRA = 1), the impact is moderated to –0.06 percent (vs. +0.02 percent in C&S).

### Economic interpretation and literature linkage
- Opposing developments in local currency versus hard currency flows:
  - Findings align with literature highlighting "original sin redux": borrowing in local currency mitigates borrower currency mismatch but transfers mismatch to lenders, exposing foreign investors to currency risk (Carstens and Shin 2019; BIS 2019).
  - During risk aversion episodes, weakening local currencies and domestic asset sell-offs prompt foreign investors to reduce exposure, triggering outflows.
  - Increased foreign currency hedging can exert additional pressure on exchange rates and funding costs even absent outright outflows.
  - Emerging market exchange rates can amplify portfolio adjustments due to pro-cyclicality: local currency spreads and exchange rates move in lockstep.
  - Conclusion: local currency issuance aids sovereigns in developing domestic markets but entails risks when a large share of investors are opportunistic and concentrated foreign investors.

*Sources: Bloomberg Finance L.P.; J.P. Morgan; authors’ calculations.*

---


_Source: https://www.imf.org/-/media/files/publications/gfs-notes/2021/english/gfsnea2021005.pdf_