What COVID-19 revealed about the resilience of bond funds

Introduction

Open-ended fixed-income mutual funds with large holdings of corporate bonds (bond funds) play a growing role in financing the Canadian corporate sector. They now hold around 23 percent of Canadian corporate bonds denominated in Canadian dollars compared with 12 percent in 2007. Bond funds offer daily redemptions to investors, but they hold assets that may be difficult to sell on short notice. If many investors were to redeem simultaneously, the funds might be forced to quickly sell corporate bonds, potentially decreasing liquidity in the bond market (Arora et al. 2019b; Bank of Canada 2019).

COVID‑19 represents a real-life test for bond funds (Falato, Goldstein and Hortaçsu 2020). Concerns about the economic impact of the pandemic created a shock wave in financial markets in March 2020. Spreads of Canadian corporate bonds widened significantly, causing the value of bond fund assets to fall. A large share of investors reacted by exiting these funds to raise cash. Net redemptions reached $14 billion in March, amounting to around 4.5 percent of assets under management (Bank of Canada 2020).

While large, these redemptions were still substantially less than those predicted by a model simulation based on the credit spreads in March (see Figure 1 for a summary of the model simulation compared with what happened in March).

The difference between the predicted and observed redemptions in March can be explained—at least partly. We believe that the Bank of Canada’s liquidity and asset purchase facilities helped calm markets and limit investor redemptions. Fund managers also played a role in preventing large redemptions by intensifying their relationship efforts with investors (i.e., they had regular conversations about investment decisions). Some fund managers also charged higher fees to investors who exited the fund. The higher fees were deemed necessary because the cost of providing liquidity in this market rose significantly in March.

Evidence suggests that most fund managers met the demand for redemptions with cash and other liquid assets. Securities regulators also gave fund managers additional flexibility to use borrowing to manage demand for redemptions. Available data indicate that, on average, cash holdings of bond funds declined from 4.2 to 3 percent of assets under management in the quarter ending in March.

Overall, the combined actions of fund managers and authorities helped prevent funds from selling bonds in a market undergoing severe liquidity strains, which would have amplified the adverse conditions of market liquidity (see Gravelle 2020 for an explanation of market liquidity and how conditions evolved during the crisis). If bond funds face another wave of large redemptions, they may be more vulnerable because they have already used part of their cash buffers. By rapidly rebuilding those buffers, bond funds can help avoid future forced sales of assets that are less liquid. The Bank will continue to monitor these funds and how they can affect fixed-income market liquidity.

Figure 1: Transmission of the COVID-19 shock on bond funds in March—observed vs. model simulation

Transmission of the COVID-19 shock on bond funds in March—observed vs. model simulation
Transmission of the COVID-19 shock on bond funds in March—observed vs. model simulation

A COVID‑19 stress simulation

We quantify how the rapid increase in Canadian credit spreads observed in March can affect redemptions of bond funds and their liquidity positions using Ceto, our mutual fund stress test model (Arora et al. 2019a).

Table 1 describes the features of the COVID-19 scenario considered in our simulation. The shock consists of movements observed in March 2020 of credit spreads across categories of credit quality. This credit spread shock is combined with the Bank of Canada’s 150-basis-point cut to the policy rate in March. The resulting overall interest rate shock is applied on a one-month horizon.

Table 1: Shock to credit spreads across the credit curve

Credit quality COVID‑19 simulation (basis points) 2007–09 global financial crisis (basis points)
AAA -100 -100
AA +20 0
A +50 +50
BBB +150 +100
BB +225 +600
B +450 +1,000
Below B +650 +2,000

While the widening of credit spreads was less acute for riskier bonds in March 2020 than during the 2007–09 global financial crisis, it was more pronounced for bonds in the AA and BBB rating segments. This partly reflects the sharp and broad-based repricing of assets caused by COVID‑19. This observation is important because, on average, around 50 percent of Canadian bond fund portfolios are securities rated AA or BBB (see Table A-1 in the Appendix).

Bond fund sample

We consider open-ended mutual funds with large holdings of Canadian corporate bonds that experienced at least 1 percent of net redemptions in March.

Our sample consists of 188 bond funds with total holdings of $323 billion in assets under management.1 Arora et al. (2019a) present information about the underlying data and explain in further detail how we identify bond funds. On average across bond funds, holdings of Canadian corporate bonds represented 44 percent of the funds’ portfolio in February 2020 (Table 2). Table A-1 in the Appendix provides the composition of bond holdings by credit ratings.

Table 2: Asset allocation of bond funds before the COVID‑19 shock

Asset allocation (%)
Cash and equivalents 4.3
Equity 12.2
Government bonds 32.4
Corporate bonds 44.3
Other 6.7

More investor redemptions after poor performance

The rapid increase in credit spreads caused by COVID‑19 led fixed-income portfolios to lose value. In March, bond funds incurred average portfolio losses of 3.4 percent, which is very close to our model prediction (3.7 percent).2

It is well documented that investors in mutual funds tend to redeem shares in response to portfolio losses. March 2020 was no exception. In fact, investors redeemed a record $14.4 billion from bond funds in March, amounting to 4.5 percent of assets under management. This was the largest monthly redemption faced by bond funds.3 Nevertheless, this is much less than what our model predicts (Table 3). The historical relationships between fund performance and redemptions suggest that the portfolio losses observed in March should have led to redemptions amounting to 9.5 percent ($30.7 billion with a 95 percent confidence interval of ± $2.9 billion).

Table 3: Bond fund redemptions

March 2020 Model simulation
Net redemptions ($ billions) 14.4 30.7
Assets under management, February ($ billions) 323 323
Redemptions as a percent of assets (%) 4.5 9.5

The difference between modelled and observed redemptions in March can be explained, at least in part, by the Bank’s establishment and expansion of liquidity and asset purchase facilities. We believe these facilities helped calm markets and limit investor redemptions. Chart 1 shows that the bulk of investor redemptions took place before the Bank announced facilities to support the functioning of funding markets important to corporations.

Chart 1: The bulk of net redemptions occured before the announcement of the Bank's facilities

Notes: Vertical lines are the dates the programs/facilities were announced. For more details on the different programs, see the Commercial Paper Purchase Program (CPPP), Contingent Term Repo Facility (CTRF) and Corporate Bond Purchase Program (CBPP). The chart is based on the subsample of bond funds that report daily flows (approximately 70 percent of our sample by assets under management [AUM]). Full sample and subsample monthly outflows are similar.
Sources: Morningstar Direct and Bank of Canada calculations Last observation: July 7, 2020

Moreover, fund managers played a role in preventing larger redemptions by intensifying their relationship management activities with both retail and institutional investors. Discussions with market participants indicate that intent to redeem shares was initially higher than the value of actual redemptions observed in March. This suggests that relationship management efforts were effective in containing redemptions. For example, because the cost of providing liquidity in this market rose significantly in March, some fund managers charged higher fees to investors who exited the fund. That way, higher exit fees created disincentives for investors (who remain in the funds) to redeem ahead of others. These mitigation measures were greater than what is normally seen. As a result, they are likely not captured in our model.

How bond funds met investor redemptions in March

When investors redeem shares, fund managers must choose a mix of cash, equity and bond sales to honour their commitments and provide cash to the investors.

Previous Bank of Canada staff analytical notes (Arora and Ouellet Leblanc 2018; Arora, Fan and Ouellet Leblanc 2019) find that bond fund managers typically use liquid asset holdings (i.e., cash and equivalents, equity and government bonds). But when they expect to continue to receive redemption requests, these managers will also sell less-liquid assets (corporate bonds) to maintain a minimum level of liquidity in their portfolios. This liquidity management strategy, called vertical slicing, could have been expected in March when stress in financial markets and redemptions surged.

But insights gathered through market intelligence suggest that most fund managers met the demand for redemptions in March with cash and other liquid assets. In doing so, fund managers minimized corporate bond sales in a market that was under severe market liquidity strains (see Chart 4 in the 2020 Financial System Review). This is validated by portfolio data: cash holdings of bond funds declined from 4.3 to 3 percent of assets under management in the quarter ending in March.

Overall, fund managers were able to manage their liquidity needs during the COVID‑19 shock in March. This highlights the importance of sound liquidity management by investment funds, an area that has been the focus of recent regulatory reforms.4

The results if redemptions had been twice as large, as implied by our model

Our model simulation predicted redemptions more than twice the size of what we observed in March ($31 billion versus $14 billion; Table 3). To assess whether fund managers could have faced these larger redemptions without having to sell less-liquid assets (thus adding stress to already-strained liquidity conditions), we simulate our model assuming that fund managers use their liquid holdings (cash, equity and government bonds) to meet redemptions.

Our simulation shows that, if funds had used liquid holdings to meet the $31 billion in model-implied redemptions, they would have experienced a $6.5 billion decline in cash, from $14.0 billion to $7.5 billion (Table 4). Despite the higher redemptions, the average share of cash holdings at the end of March would not have declined much more than what is observed (2.6 versus 3 percent).

Table 4: The impact of the COVID‑19 shock on asset holdings of bond funds

$ billions (% of assets under management) Before COVID‑19 After COVID‑19: observed After COVID‑19: model simulation
Cash and equivalents 14.0 (4.3) 8.4 (3.0) 7.5 (2.6)
Equity 36.5 (11.3) 23.1 (8.2) 34.6 (11.8)
Government bonds 107.4 (33.2) 104.0 (36.9) 90.8 (31.0)
Corporate bonds 144.3 (44.7) 133.1 (47.2) 136.4 (46.7)
Other 20.9 (6.5) 13.4 (4.8) 23.0 (7.9)

But this is true in aggregate only: this result masks important differences across funds. If redemptions had reached the level predicted by the model, around 70 percent of funds would have exhausted their cash buffers to meet redemptions. In contrast, available data suggest that 9 percent of funds had drawn from their cash buffers entirely at the end of March.

Unsurprisingly, Table 4 also shows that the use of liquid holdings to meet redemptions has led to a decrease in the share of liquid holdings and an increase in the share of less-liquid assets (corporate bonds). This implies that bond funds could find it difficult to manage another wave of large redemptions without having to sell less-liquid assets.

The cash positions of bond funds following the COVID‑19 shock

As just mentioned, bond funds’ share of cash holdings as a percentage of assets under management (their liquidity ratio) declined to 3 percent, on average, in March from 4.3 percent in December. Fund-level data confirm that this trend toward a smaller share of cash holdings is true for the bulk of bond funds. Chart 2 illustrates the shift in the distribution of the liquidity ratio across funds between December 2019 and March 2020. Of the 188 bond funds in our sample, 112 had reduced their liquidity ratio at the end of March.

Chart 2: In March, 60 percent of bond funds experienced a decline of their cash holdings

Sources: Morningstar Direct and Bank of Canada calculationsLast observation: March 2020

Bond funds are thus potentially more vulnerable should they face another wave of large redemptions. Chart 3 compares the liquidity ratio with the two worst monthly redemptions observed since 2007 (October 2008 and March 2020). The current cash holdings would be enough to cover monthly redemptions equivalent to those observed in October 2008 but not those in March 2020.

Chart 3: The average liquidity ratio has fallen to its lowest level since 2007

Note: The liquidity ratio is cash and cash equivalents as a percentage of assets under management.
All dates are year-end except 2020, which is data for June 2020.
Sources: Morningstar Direct and Bank of Canada calculations Last observation: June 2020

Although redemptions have stopped since April, lower liquidity ratios raise concerns about the ability of bond funds to meet potentially large redemption requests. If redemptions were to accelerate in the near term, fund managers with low cash buffers may be forced to sell less-liquid assets, thus amplifying asset price declines and contributing to deterioration in market liquidity. By rapidly rebuilding their cash buffers, fund managers can help avoid such undesirable situations.

Other measures have also been taken to manage large liquidity demand from investors in the near term. Securities regulators gave bond fund managers additional flexibility to use borrowing to deal with large redemption requests.5 In addition, the Bank’s recent activation of the Contingent Term Repo Facility could provide liquidity to market participants outside the traditional banking sector, such as fund managers.

Conclusion

At this stage, Canadian bond funds have remained resilient to COVID‑19 challenges. The liquidity management strategies of fund managers, supported by policy measures, have helped bond funds limit the increase in redemptions and manage these redemptions without adding stress to the financial system.

Nevertheless, the surge in redemptions in March left these funds with lower cash buffers, which could make them more vulnerable to another round of large redemptions. The Bank of Canada will closely monitor how bond funds are strengthening the resilience of their liquidity strategies by rebuilding their liquidity buffers to manage risks in the future.

Appendix

Table A-1 shows the average fund allocation to fixed-income securities by credit rating.

Table A-1: Credit allocation of bond funds

Credit quality Allocation (%)
AAA 25.8
AA 30.4
A 16.7
BBB 18.1
BB 5.1
B 3.0
Below B 0.9

  1. 1. Of the 188 bond funds in our sample, 50 are mixed funds with large holdings of Canadian corporate debt. Their total size accounts for 25 percent of our sample ($81 billion). Due to their investment mandate, mixed funds do not hold only bonds, which explains why our sample holds a small share of equities (Table 2).[]
  2. 2. The model captures the historical relationships that exist between interest rates and the portfolio performance of individual funds.[]
  3. 3. During the 2007–09 global financial crisis, the largest monthly redemption occurred in October 2008 and represented 2.9 percent of assets under management.[]
  4. 4. In recent years, global and domestic authorities have strengthened regulatory guidance related to liquidity management, leverage and concentration limits. See, for example, Financial Stability Board, Policy Recommendations to Address Structural Vulnerabilities from Asset Management Activities (January 12, 2017). See also IOSCO, Recommendations for Liquidity Risk Management for Collective Investment Schemes (February 2018). In the Canadian context, an example of securities regulations can be found in “OSC Staff Notice 81-727 Report on Staff’s Continuous Disclosure Review of Mutual Fund Practices Relating to Portfolio Liquidity.”[]
  5. 5. Specifically, the Canadian Securities Administrators (CSA) will allow a temporary increase of borrowing limit (from 5 to 10 percent of a fund’s net asset value) to accommodate potential redemption requests. See the CSA press release, “Canadian Securities Regulators Temporarily Increase Short-Term Borrowing Limits for Mutual Funds Investing In Fixed Income,” April 17, 2020.[]

References

  1. Arora, R., G. Bédard-Pagé, G. Ouellet Leblanc and R. Shotlander. 2019a. “Bond Funds and Fixed-Income Market Liquidity: A Stress-Testing Approach.” Bank of Canada Technical Report No. 115.
  2. Arora, R., G. Bédard-Pagé, G. Ouellet Leblanc and R. Shotlander. 2019b. “Could Canadian Bond Funds Add Stress to the Financial System?” Bank of Canada Staff Analytical Note No. 2019-9.
  3. Arora, R., C. Fan and G. Ouellet Leblanc. 2019. “Liquidity Management of Canadian Corporate Bond Mutual Funds: A Machine Learning Approach.” Bank of Canada Staff Analytical Note No. 2019-7.
  4. Arora, R. and G. Ouellet Leblanc. 2018. “How Do Canadian Corporate Bond Mutual Funds Meet Investor Redemptions?” Bank of Canada Staff Analytical Note No. 2018-14.
  5. Bank of Canada. 2020. Financial System Review (May).
  6. Bank of Canada. 2019. Financial System Review (May).
  7. Falato, A., I. Goldstein, A. Hortaçsu. 2020. “Financial Fragility in the COVID-19 Crisis: The Case of Investment Funds in Corporate Bond Markets.” National Bureau of Economic Research Working Paper No. 27559.
  8. Gravelle T. 2020. “Keeping markets liquid.” The Economy, Plain and Simple. Bank of Canada (June 10).

Acknowledgments

We thank Rohan Arora, Don Coletti, Jean-Philippe Dion, Joshua Slive and Virginie Traclet for helpful comments and suggestions. Finally, we are grateful to Alison Arnot and Meredith Fraser-Ohman for editorial assistance.

Disclaimer

Bank of Canada staff analytical notes are short articles that focus on topical issues relevant to the current economic and financial context, produced independently from the Bank’s Governing Council. This work may support or challenge prevailing policy orthodoxy. Therefore, the views expressed in this note are solely those of the authors and may differ from official Bank of Canada views. No responsibility for them should be attributed to the Bank.

JEL Code(s): G, G1, G2, G20, G23