Evaluating the portfolio balance effects of the Government of Canada Bond Purchase Program on the Canadian yield curve

The Bank of Canada’s response to the COVID‑19 pandemic

In response to the severe market dysfunction and the sharp increase in demand for liquidity that happened at the onset of the COVID‑19 pandemic, the Bank of Canada undertook several extraordinary measures to stabilize financial markets and support the Canadian economy. The Bank decreased its policy interest rate to the effective lower bound. In addition, the Bank launched several liquidity and asset purchase programs in March 2020 that targeted various segments of the financial market. These programs included purchases of:

• government bonds
• Canada Mortgage Bonds
• corporate bonds
• commercial paper
• bankers’ acceptances

Unprecedented in scale, these asset purchase programs represented 23.6% of Canada’s 2019 nominal gross domestic product (GDP).1

One program in particular—the Government of Canada Bond Purchase Program (GBPP)—played a pivotal role in the Bank’s pandemic response, accounting for almost 80% of assets purchased.

Initially, the GBPP aimed to address the severe market dysfunction and to restore liquidity in the government bond market by purchasing at least $5 billion of Government of Canada (GoC) bonds per week. However, by July 2020, as market conditions normalized, the GBPP transitioned to a quantitative easing (QE) program to provide additional monetary stimulus.

The shift to QE was designed to decrease long-term borrowing costs in the economy because purchasing GoC bonds of a given maturity bids up their price and thereby reduces the interest rate that the investors earn from holding bonds. This lower interest rate, in turn, transmits to mortgages and corporate loans. This stimulates more borrowing and spending and, consequently, supports economic recovery and helps the Bank achieve its inflation target.2

I focus on the effect the GBPP had on the interest rate for GoC bonds. This is because the yields on these bonds often serve as a benchmark for the returns required for other, riskier assets such as corporate bonds, mortgages and business loans. Indeed, my estimates indicate that, at its peak, the GBPP reduced the Canadian 10‑year and 5‑year zero-coupon yields by at least 84 basis points (bps) and 52 bps, respectively. Importantly, the effects of the GBPP on long-term yields are persistent, still reducing the 10‑year and 5‑year zero-coupon yields by 51 bps and 32 bps, respectively, by the end of March 2024.

The effect of central banks’ large-scale bond purchases on long-term interest rates

Central bank bond purchases decrease long-term interest rates primarily through a mechanism known as the portfolio balance channel.3 The main idea behind this channel is that investors do not view all assets as perfect substitutes. This is due to various factors, such as differences in liquidity, risk characteristics, tax treatment, regulatory constraints and institutional preferences.

This imperfect substitutability means that changes in the net supply of a security can affect the prices and yields of other assets as investors rebalance their portfolios.

Imperfect substitutability among government bonds arises because these bonds do not all respond the same way to movements in the interest rate. Specifically, when interest rates increase, investors are concerned about the risk of capital losses on their bond holdings. Long- and short-term bonds are therefore considered imperfect substitutes because long-term bond prices are more sensitive to changes in the interest rate. In other words, when interest rates rise, the prices of long-term bonds fall more than those of short-term bonds; the opposite is true when interest rates decline.

As a result, investors typically demand higher yields on long-term bonds to compensate for the greater interest rate risk they have compared with short-term bonds. Under the GBPP, the Bank purchased long-term government bonds using short-term assets, such as settlement balances, essentially swapping the long-term assets for short-term ones in the portfolios of private investors. This reduces private investors’ exposure to interest rate risk and puts downward pressure on long-term interest rates.

The average maturity of outstanding Government of Canada debt

To illustrate the amount of interest rate risk the GBPP removed from the market, I compute the reduction in the average maturity of the outstanding GoC marketable debt resulting from the GBPP. Specifically, I collect data on bond characteristics—issue date, coupon rate, maturity date and face value outstanding—for every nominal GoC bond outstanding between January 2018 and March 2024 as well as data on the Bank’s holdings of each of these bonds.

Importantly, the Bank also acquires GoC bonds to offset government deposits and the issuance of bank notes. Acquired primarily through non‐competitive bids at government securities auctions, these bonds represent liabilities on the Bank’s balance sheet.4 These acquisitions are referred to as purchases under normal course. To isolate the effects of the GBPP, in my dataset I distinguish between holdings of GoC bonds acquired:

• on the primary market under normal course5
• on the secondary market due to the GBPP

Chart 1 shows three measures of the average maturity of the outstanding GoC marketable debt.6 The blue line represents the average maturity of the outstanding GoC marketable debt. The yellow line represents the average maturity of the outstanding GoC marketable debt that consolidates the Bank’s holdings of GoC purchased under normal course. Finally, the green line displays the average maturity of the outstanding GoC marketable debt that consolidates the Bank’s total holdings of GoC debt.7 The difference between the average maturity of the outstanding GoC debt consolidates the Bank’s holdings due to its normal operations (yellow line) and the total holdings (green line) isolates the effect of the GBPP.

As the pandemic progressed, the average maturity of the outstanding GoC debt gradually began returning to pre-pandemic levels of around 5 years, from a level of 3.7 years in June 2020. This reflects the government’s gradual shift back toward issuing long-term bonds once market conditions had stabilized and the initial liquidity crunch had eased. At the same time, the Bank also shifted its focus, changing the purpose of its extraordinary actions from focusing on restoring market functioning to serving as a QE tool to provide additional monetary stimulus while the policy rate was constrained by the effective lower bound. Effectively, as the Bank ramped up its GBPP purchases starting in March 2020, the program significantly affected the interest rate risk that private investors were exposed to. By November 2021—when the Bank’s GoC bond holdings under the GBPP were at their peak—the weighted average maturity of GoC bonds held by private investors had decreased by about 1.4 years.8

The Bank started normalizing its balance sheet by first stopping net new purchases of GoC bonds in October 2021 when it moved from QE to the reinvestment phase.9 In April 2022, the Bank stopped all purchases of GoC bonds.10 As a result, the gap between the yellow and green lines in Chart 1 started to close. By the end of March 2024, the effect of the GBPP on the average maturity of the GoC outstanding debt was still 0.87 years.11

Impact on 10-year interest rates

I build a novel dynamic portfolio balance model of the term structure of interest rates for Canada to map how a 1.4 year decline in the average maturity of the portfolio of GoC debt that fixed-income investors must hold affects the Canadian yield curve. For further details, see Diez de los Rios (2024).

As in Greenwood and Vayanos (2014) and Vayanos and Vila (2021), marginal holders of long-term government securities in my model are risk-averse fixed-income investors who dislike the risk of unexpected capital losses on their bond portfolio when interest rates increase. For this reason, these investors are willing to absorb an increase in the supply of long-term nominal bonds only when compensated by a risk premium in the form of an increase in long-term interest rates relative to expected future short-term interest rates. By decreasing the amount of long-term debt in the market, central banks can reduce the risk fixed-income investors are exposed to, thus lowering the risk premium component of long-term interest rates. In other words, the amount of long-term government debt that these financial intermediaries need to hold can affect bond risk premiums and, consequently, bond yields.

Using my model, I can calculate how much compensation private fixed-income investors require—in the form of a bond term premium—for holding a particular portfolio of GoC bonds (Chart 2).12 Reducing by 1.4 years the weighted average maturity of the portfolio of GoC debt that fixed-income investors must hold decreases the term premium component of GoC bond yields. This effect relies on the assumed value for the persistence of the QE program and on the risk aversion of fixed-income investors. Consistent with the work of King (2019) for the United States, I start by considering the effects of a QE program with a half-life of 4.5 years for my baseline case.13 In addition, I calibrate the risk aversion of the fixed-income investors such that, when I estimate the model using US data, I am able to match the empirical evidence on the impact on the yield curve of the US Federal Reserve’s pre-pandemic QE programs.14

Using these assumptions, my model suggests that the Bank’s GBPP had portfolio balance effects of 84 bps on the 10‑year yield. These results are in line with those in Azizova, Witmer and Zhang (2024), who follow a different approach. They use an event study to look at how much Canadian bond yields changed when purchases under the GBPP were announced, and they find that the GBPP had an impact on 10‑year government bond yields of about 20 bps. However, since some investors anticipated that the Bank would purchase GoC bonds even before the GBPP was announced, these authors estimate that that the GBPP effect on 10‑year bond yields was almost 80 bps.15

The portfolio balance effects are more pronounced on long-term yields. For example, the portfolio balance effects of the Bank’s GBPP on the 2‑year yield were 20 bps. In comparison, the effects on the 5‑year and 10‑year yields were 52 bps and 84 bps, respectively.

However, this effect is nonlinear because investors are forward looking: they understand that the QE shock is temporary and that its influence will wane over the life of the bond. They therefore require less additional compensation for holding bonds with increasing maturities.

To check the robustness of my results, I use two additional values of the persistence of the QE program proposed by Vayanos and Vila (2021). I find that when the QE program is more persistent (half-life of 6.9 years), the impact on the 10‑year and 5‑year bond yields increases to 98 bps and 56 bps, respectively. When the QE program is less persistent (half-life of 3.5 years), the GBPP’s impact on the 10‑year and 5‑year bond yields declines to 77 bps and 50 bps, respectively. Importantly, as the persistence of the QE program increases, the relationship between the portfolio balance effect on yields and the maturity of the bond becomes more linear.

Finally, to provide a sense of the persistence of the effects of the GBPP on Canadian yields, I also calculate the effect of the 0.87‑year reduction on the average maturity outstanding of GoC debt due to the Bank’s holdings of GoC debt purchased under the GBPP as at the end of March 2024. My model suggests that a 0.87‑year reduction in the average maturity of the outstanding GoC debt has a portfolio balance impact of 51 bps on the 10‑year yield and 32 bps on the 5‑year yield—approximately 60% of the peak impact.

Conclusion

In response to the increase in the demand for liquidity and market dysfunction from the COVID‑19 pandemic, the Bank undertook extensive measures to stabilize financial markets and support the Canadian economy. Of these programs, the GBPP was critical to the Bank’s response to the pandemic because the Bank transitioned it to a QE program to provide additional monetary stimulus.

My calculations suggest that GBPP was effective in lowering the Canadian 10‑year and 5‑year zero-coupon yields by 84 bps and 52 bps, respectively.

References

Azizova C., J. Witmer, and X. Zhang. 2024. “Assessing the Impact of the Bank of Canada’s Government Bond Purchases.” Bank of Canada Staff Discussion Paper 2024-5.

Bank of Canada. 2021. “Statement of Policy Governing the Acquisition and Management of Financial Assets for the Bank of Canada’s Balance Sheet.”

Bank of Canada. 2021. “Bank of Canada Maintains Policy Rate and Forward Guidance, Ends Quantitative Easing.” Press release (October 27).

Bank of Canada. 2022. “Bank of Canada Increases Policy Interest Rate by 50 Basis Points, Begins Quantitative Tightening.” Press release (April 13).

Bauer, M. D. and G. D. Rudebusch. 2014. “The Signaling Channel for Federal Reserve Bond Purchases.” International Journal of Central Banking 10 (3): 233–289.

Beaudry, P. 2020. “Our Quantitative Easing Operations: Looking Under the Hood.” Speech delivered virtually to the Greater Moncton Chamber of Commerce, the Fredericton Chamber of Commerce and the Saint John Region Chamber of Commerce, New Brunswick, December 10.

Diez de los Rios, A. 2024. “Estimating the Portfolio Balance Effects of the Bank of Canada's Government Bond Purchase Program.” Bank of Canada Staff Working Paper 2024-34

Diez de los Rios, A. forthcoming. “A Portfolio-Balance Model of Inflation and Yield Curve Determination.” Review of Asset Pricing Studies.

Gagnon, J., M. Raskin, J. Remache and B. Sack. 2011. “The Financial Market Effects of the Federal Reserve’s Large-Scale Asset Purchases.” International Journal of Central Banking 7 (1): 3–43.

Greenwood, R., S. G. Hanson, J. S. Rudolph and L. H. Summers. 2016. “The Optimal Maturity of Government Debt.” In The $13 Trillion Question: How America Manages Its Debt, edited by David Wessel, 1–41. Washington DC: Brookings Institution Press.

Greenwood, R., S. G. Hanson and D. Vayanos. 2015. “Forward Guidance in the Yield Curve: Short Rates versus Bond Supply.” National Bureau of Economic Research Working Paper No. 21750.

Greenwood, R. and D. Vayanos. 2014. “Bond Supply and Excess Bond Returns.” Review of Financial Studies 27 (3): 663–713.

Johnson G. 2023. “A Review of the Bank of Canada’s Market Operations Related to COVID‑19.” Bank of Canada Staff Discussion Paper No. 2023-6.

Joyce, M., A. Lasaosa, I. Stevens and M. Tong. 2011. “The Financial Market Impact of Quantitative Easing in the United Kingdom.” International Journal of Central Banking 7 (3): 113–161.

King, T. B. 2019. “Expectation and Duration at the Effective Lower Bound.” Journal of Financial Economics 134 (3): 736–760.

Kozicki, S. 2024. “Exceptional Policies for an Exceptional Time: From Quantitative Easing to Quantitative Tightening.” Speech to Canadian Association of Business Economics, Ottawa, Ontario, June 13.

Krishnamurthy, A. and A. Vissing-Jorgensen. 2011. “The Effects of Quantitative Easing on Interest Rates: Channels and Implications for Policy.” Brookings Papers on Economic Activity (Fall): 215–265.

Tombe, T. 2023. “From Bonds to Banknotes: Central Banking and Public Finances in Canada.” Canadian Tax Journal 71 (3): 825–852.

Vayanos, D. and J. Vila. 2021. “A Preferred-Habitat Model of the Term Structure of Interest Rates.” Econometrica 89 (1): 77–112.

Zhang, X. 2021. “Evaluating the Effects of Forward Guidance and Large-Scale Asset Purchases.” Bank of Canada Staff Working Paper No. 2021-54.

Acknowledgements

I would like to thank James Chapman, Toni Gravelle, Serdar Kabaca, Sharon Kozicki, Gitanjali Kumar, Stéphane Lavoie, Miguel Molico, Jonathan Witmer and seminar participants at the Bank of Canada for their suggestions. Ayesha Siddiqua and Amanda Wang provided excellent research assistance. Any remaining errors are my own. The views expressed in this paper are those of the author and do not necessarily reflect those of the Bank of Canada.

Endnotes