Mutual Funds and Government Bonds Portfolio Reallocation

 The sources detail several crucial findings regarding US mutual funds' demand for safe government bonds, challenging the traditional assumption that demand for safe assets is inelastic to yields. The paper specifically investigates how US mutual funds, a key class of investors in global safe assets, actively adjust their portfolios based on cross-currency yield differentials.

The main research findings concerning US mutual funds' investment behavior, situated within the context of safe government bonds and returns, are centered around yield-elasticity, currency hedging, and reactions to global financial conditions.

1. Yield-Elastic Demand and Unhedged Returns

A primary finding is that US-based fund managers actively rebalance their portfolios towards government bonds that offer relatively higher returns on an unhedged basis (i.e., without hedging currency risk).

• Magnitude of Effect: The size of this effect is significant and large. A change by one percentage point in the unhedged excess return typically leads to a change in portfolio share of around one percentage point on average across several currencies.
• Target Currencies (Unhedged): This rebalancing behavior is particularly notable towards euro-denominated securities.
• Capital Flows Impact: This portfolio adjustment has substantial implications for capital flows. For instance, a one percentage point increase in the excess return of euro area government debt securities would trigger capital flows from the United States to the euro area economies issuing highly-rated debt securities in the magnitude of $300 million. This capital flow represents about 2 percent of total quarterly foreign flows into highly-rated euro area government debt securities.
• Search for Yield: This strong reaction suggests a "search for yield" behavior among safe government bonds.

2. Hedged Returns and Covered Interest Parity (CIP) Deviations

The study also provides novel insights by examining the reaction of mutual funds to currency hedged excess returns, which are relevant when asset managers use derivatives to neutralize exchange rate fluctuations. CIP deviations occur when the dollar return does not equal the hedged return from investing in another currency, creating a wedge.

• Exploiting CIP Deviations: US mutual funds actively rebalance their portfolios towards currencies, such as the Japanese yen, that display large CIP deviations and offer higher returns on a hedged basis compared to US Treasuries.
• Role as Liquidity Providers: These results indicate that US mutual funds exploit the advantage provided by their role as liquidity providers in the market for forward dollars, where frictions lead to CIP deviations.
• Specific Example (JPY): Government debt denominated in the Japanese yen typically offers the highest and least volatile hedged excess return among the sampled government debt securities. A one standard deviation change in Japanese yen hedged excess returns is estimated to trigger a reallocation of approximately 150 basis points in its portfolio share.
• Differences in Reaction: There are significant differences in how funds react to returns based on whether they are unhedged or hedged. For most currencies, the sensitivity of portfolio shares to hedged returns is lower, which is consistent with the low predictability of hedged excess returns and potentially low elasticity of hedging demand to exchange rates.

3. Influence of Global Risk and Monetary Policy

The sources investigate the role of financial and monetary conditions, characterizing the behavior of safe assets beyond just returns.

• Response to Global Financial Risk (Flight to Safety): When global financial risk is on the rise, often proxied by the VIX index, US mutual fund managers repatriate their investments towards US government debt securities. This reallocation primarily occurs at the expense of euro-denominated securities. This retrenchment is consistent with a flight-to-safety argument, recognizing US Treasuries as the premier global safe asset.
• Response to Low US Policy Rates (Search for Yield): The evidence suggests that US professional investors search for yield in the safe government bond context when US monetary policy rates are low. When the Federal Reserve's policy rate is low, the sensitivity of the euro's currency share to excess returns in the US mutual funds' sovereign portfolio is more elevated.
• Response to High US Policy Rates (Safe Haven Flow): Conversely, times of tight monetary policy in the US are associated with higher portfolio shares for the Japanese yen and the Swiss franc, suggesting a flight to safety towards these currencies.

4. Portfolio Adjustment Dynamics

The study also provides insights into how the mutual funds adjust their portfolios over time:

• Portfolio Frictions (Inertia): There is strong evidence of portfolio frictions, reflected by positive and statistically significant coefficients associated with lagged currency shares, indicating slow portfolio adjustment.
• Valuation Effects: Generally, valuation effects driven by bond price and exchange rate movements do not influence currency shares, implying that fund managers actively offset these passive changes through rebalancing. However, fluctuations in the exchange rate of the Japanese yen and the euro appear to have an impact. For the euro, the data suggests a currency momentum strategy where managers actively increase exposure after the euro appreciates.

These central findings underscore that, contrary to traditional views, the demand by US mutual funds for safe government bonds is strongly elastic to currency yield differentials, whether those yields are accessed on an unhedged basis (e.g., in the euro area) or via exploiting arbitrage failures on a hedged basis (e.g., with the Japanese yen).

---

The behavior of US mutual funds in this context is similar to a global water system where pipes (mutual funds) quickly divert water (capital) towards pressure points offering the highest flow (excess returns), whether those pressure points are naturally high (unhedged yield) or created artificially through market imbalances (hedged CIP deviations). When a global storm hits (rising risk aversion), the system actively redirects flow back to the largest, safest reservoir (US Treasuries).

The sources provide a detailed context for understanding safe assets and illuminate specific drivers of demand for these assets, particularly among US mutual funds, often contradicting the traditional view regarding yield-elasticity.

I. Context and Characteristics of Safe Assets

Safe assets are central to global finance, and their characteristics and the drivers of their demand have been a focus of interest, especially since the global financial crisis of 2008.

Definition and Scarcity: Safe assets generally refer to government debt securities issued by major advanced economies that possess a reserve currency status and maintain the highest credit rating (double A or higher according to Standard & Poor's). Since 2008, a scarcity of these assets has emerged, causing a dramatic decline in their yields. The continuing rise in geopolitical risk (such as the war in Ukraine and tensions in the Middle East) makes understanding these demand drivers even more crucial, given the potential long-term consequences for the international monetary system.

Key Characteristics: Beyond their high credit ratings, safe assets possess several valued properties:

• They command a premium for safety and liquidity.
• They are information insensitive.
• They exhibit a negative market beta, meaning they appreciate when global risk aversion rises. They are described as a "good friend," valuable and liquid when one needs them.
• They are desirable because they attract investor demand and maintain market value and liquidity during periods of financial stress.
• US Treasury debt is widely considered the world's premier safe asset.

Challenging the Traditional View of Demand: A main theoretical feature of safe assets is the assumption of relatively low elasticity of their demand with respect to yields. However, this paper contributes to mounting evidence showing this is not always the case. The study demonstrates that cross-currency yield differentials in the sovereign bond market of high-rating issuers significantly influence the appeal of currencies for US mutual funds, thus shaping the overall demand for global safe assets.

II. Mutual Fund Demand Drivers (Returns)

The demand for safe government bonds by US mutual funds is found to be highly responsive to returns, suggesting an active "search for yield" behavior.

Unhedged Excess Returns (Direct Yield Differential): US-based fund managers actively rebalance their portfolios towards government bonds that offer relatively higher returns on an unhedged basis (i.e., without hedging currency risk).

• This "search for yield" is particularly noticeable towards euro-denominated securities issued by highly-rated euro area economies (Austria, Belgium, France, Germany, and the Netherlands).
• The magnitude of this effect is large: a change of one percentage point in the unhedged excess return leads to an average change in portfolio share of around one percentage point across several currencies.
• This portfolio adjustment triggers sizeable capital flows; for example, a one percentage point increase in the excess return of euro area debt would trigger flows of approximately $300 million from the US to those euro area economies.

Hedged Excess Returns (Arbitrage Opportunities): US mutual funds also respond to returns calculated on a currency-hedged basis, exploiting Covered Interest Parity (CIP) deviations.

• Funds rebalance towards currencies, notably the Japanese yen, which display large CIP deviations and offer higher returns on a hedged basis compared to US Treasuries.
• This behavior suggests that US mutual funds exploit an advantage stemming from their role as liquidity providers in the market for forward dollars, where frictions lead to CIP deviations.
• The persistence and magnitude of CIP deviations affect the portfolio choice of mutual funds, which in turn drives large capital flows from the United States.

III. Mutual Fund Demand Drivers (Financial and Monetary Conditions)

The sources also provide insights into how global conditions and US monetary policy influence the demand for different safe assets.

Global Risk Aversion (Flight to Safety): In times of rising global financial risk, typically proxied by the VIX index, US mutual fund managers exhibit a classic "flight-to-safety" behavior.

• Managers repatriate their investments towards US government debt securities.
• This retrenchment primarily occurs at the expense of euro-denominated securities.
• This active reallocation towards domestic bonds reinforces the status of US Treasuries as the global safe asset par excellence.

US Monetary Policy Rates: The environment created by the Federal Reserve’s policy rate influences the sensitivity of fund managers to excess returns.

• Low US Policy Rates: When the US policy rate is low, the sensitivity of the euro's currency share to excess returns is more elevated. This strongly suggests that US professional investors engage in a search for yield abroad in the context of safe government bonds when domestic rates are low.
• High US Policy Rates: Periods of tight monetary policy in the US are associated with higher portfolio shares for the Japanese yen and the Swiss franc. This is consistent with a flight to safety behavior towards these foreign safe-haven currencies, potentially because high US policy rates signal global financial stress.

Overall, the sources portray the demand for safe assets not as static or inelastic, but as a dynamic process driven by relative currency returns (both hedged and unhedged) and contextualized by global risk appetite and US monetary policy conditions. The US mutual fund industry acts as a crucial channel, translating these differentials and risk perceptions into significant international capital flows.

The sources provide comprehensive details on the methodology and data used to investigate the demand for safe government bonds by US mutual funds, specifically focusing on the elasticity of demand to currency unhedged and hedged excess returns.

I. Data Sources and Scope

The research relies on a detailed, granular dataset covering US-domiciled mutual funds:

• Primary Dataset: The core data comes from Refinitiv Lipper, which provides detailed, fund-level panel data regarding the portfolios of US-domiciled mutual funds.
• Time Period: The analysis covers a quarterly period from 2010 Q1 to 2021 Q4. The panel used is unbalanced, as funds enter and drop out of the sample.
• Fund Selection: The analysis is restricted to fixed-income funds with an active management style. Mixed-funds (which substitute equity for bonds) are excluded.
• Total Sample Size (Initial): The initial sample included 880 funds.
• Coverage: The study uses data covering a large portion of the market, with the total Assets under Management (AuM) of the selected sample ranging from $100 billion in 2010 Q3 to $366 billion in 2021 Q3. This coverage represents about 4% to 11% of the AuM held by all US-based mutual funds investing in the fixed-income market. This size and coverage are noted as being larger than those of previous studies using portfolio-level data on fixed-income mutual funds.

II. Defining the Portfolio of Safe Assets

The study defines and constructs a specific portfolio of safe assets to analyze the investment decisions of US mutual funds:

• Sovereign Issuers: The portfolio focuses on government debt issued by major advanced economies that maintained an S&P credit rating of AA or above throughout the sample period.
• Selected Countries/Currencies: The analysis focuses on the US dollar, the euro (specifically bonds issued by Austria, Belgium, France, Germany, and the Netherlands), the Japanese yen, the pound sterling, the Swiss franc, the Australian dollar, and Canada.
• Rationale for Selection: This selection ensures the focus is on currencies with a low degree of credit risk and major international importance. It also includes currencies where Covered Interest Parity (CIP) deviations offer different signs of extra return from a US dollar investor's perspective (e.g., EUR and JPY often offering extra return, AUD typically negative) to investigate if the reaction to hedged returns changes with the sign of the CIP deviations.
• Currency Share Calculation: The share ($s_{ji,t}$) of country (currency) $j$ bonds held by fund $i$ in quarter $t$ is calculated as the ratio of the market value of country $j$ government bonds (summed over all maturities) to the market value of government bonds of all selected countries, likewise summed over all maturities.

III. Addressing Sample Bias (Home Bias)

A significant methodological step was addressing the strong home bias prevalent in US-based funds' portfolios:

• The initial sample showed an aggregate portfolio share heavily biased towards US Treasuries, accounting for more than 80% of the total portfolio.
• To focus on fund managers with a diversified international portfolio, funds that had an average portfolio share greater than 95% in any single country of interest were excluded.
• The restricted sample includes 186 funds. In this restricted sample, the US share drops substantially (averaging around 43%), aligning more closely with the theoretical International Capital Asset Pricing Model (ICAPM) benchmark, indicating a successful selection of geographically well-diversified funds.

IV. Methodology for Portfolio Reallocation and Econometric Modeling

A key component of the methodology is isolating the active decisions of fund managers from passive market movements:

1. Decomposition of Portfolio Share Change: The change in a fund's currency portfolio share ($\Delta s_{ji,t}$) is mathematically decomposed into an active reallocation component ($\Delta s_{j,A_{i,t}}$) and passive reallocation components due to bond price returns ($\Delta s_{j,P,R_{i,t}}$) and exchange rate movements ($\Delta s_{j,P,XR_{i,t}}$). The sources indicate that active reallocation dominates the overall variation in currency shares for all currencies.
2. Addressing Outliers: Outliers, which likely reflect misreported data or radical, non-return-related strategy changes (e.g., portfolio share changes of 100% in either direction), are addressed by augmenting regressions with a dummy variable that takes the value of 1 for fund-quarter observations in the top or bottom 1% by active rebalancing.
3. Baseline Specification: The main empirical approach uses panel regressions of currency shares on the lagged share, fund-level excess returns, and passive reallocation components.
   • Fixed Effects: The model includes fund fixed effects (to control for fund-specific characteristics like management style) and time fixed effects (to control for global and currency-specific time-varying factors, such as aggregate demand shocks or generalized flight-to-safety behavior).
   • Identification Strategy: The use of fund and time fixed effects helps to isolate the idiosyncratic variation in fund-level, currency-specific excess returns, which identifies the average sensitivity of portfolio shares to these returns, assuming fund-specific demand shocks do not affect prices or exchange rates.
   • Portfolio Frictions and Valuation Effects: The inclusion of the lagged share tests for portfolio frictions (slow adjustment), while coefficients on passive reallocation components determine if managers actively offset or promote valuation effects (e.g., currency momentum).
   • Standard Errors: Driscoll and Kraay (1998) standard errors are used in the regressions.

V. Measuring Excess Returns (The Main Explanatory Variable)

The variable of interest is the fund-specific excess return ($r^{ex}_{j,i,t}$), measuring the attractiveness of investing in currency $j$ relative to the fund's current sovereign portfolio.

1. Unhedged Excess Returns ($r^{ex,unh}_{j,t}$): This is the total return differential between US dollar debt and debt issued in another currency, capturing direct yield differences and currency risk.
   • Calculation is based on the yield of country $j$'s government bond in the domestic currency, averaged over four maturities (3 months, 1 year, 2 year, and 5 year).
   • It assumes future exchange rate expectations equal the current value, consistent with the exchange rate random walk hypothesis.
   • Unhedged returns display positive, significant, and large autocorrelation up to four quarters ahead for most currencies, suggesting they provide a good forward-looking signal for investors.
2. Hedged Excess Returns ($r^{ex,fwd}_{j,t}$): This measures the return when asset managers use derivatives, specifically forward contracts, to neutralize the impact of exchange rate fluctuations.
   • Hedged returns are calculated using the forward exchange rate, averaged over the same four maturities.
   • These returns are relevant because roughly 90% of US fixed-income funds with an international focus use currency forwards to manage foreign exchange exposure.
   • Hedged excess returns measure the deviation from Covered Interest Parity (CIP), which creates a wedge between the dollar return and the foreign currency hedged return.
   • In contrast to unhedged returns, hedged returns display much smaller autocorrelation coefficients (significant only up to a maximum of two quarters for a few currencies), indicating lower predictability.

The methodological robustness of these choices is also explored through various checks, including using Pooled OLS, analyzing the whole sample (which showed that the main findings were slightly biased toward zero without the restricted sample selection), testing for survivorship bias, and using an ICAPM-based selection procedure. The analysis confirms the stability of core results across most tests.