# How does the Bank of Canada’s balance sheet impact the banking system?

## Introduction

The COVID‑19 pandemic caused severe stress in fixed-income markets. In response, in April 2020, the Bank of Canada launched the Government of Canada (GoC) bond purchase program. Initially, the program focused on restoring market functioning in the GoC bond market. In July 2020, that focus shifted to providing additional monetary stimulus through quantitative easing (QE).1

Under QE, the Bank bought government bonds in exchange for settlement balances and, as a result, its balance sheet expanded.2, 3 Settlement balances (or reserves) are deposits that major Canadian banks hold at the Bank (Chu et al. 2022).

The Bank’s quantitative tightening (QT) program, which began in April 2022, is the reverse process. Through QT, the Bank allows its holdings of GoC bonds to mature and stops reinvesting the proceeds of principal and coupon repayments. As a result, the Bank’s balance sheet will shrink over time.

In this note, we describe how both QE and QT affect the balance sheets of the Bank and the overall Canadian banking system.4 We show that the direct effects on the size, composition and liquidity of the banking system’s balance sheet during QE and QT depend on who sells (during QE) or buys (during QT) GoC bonds in the financial system—banks or non-bank participants (such as households, businesses or investment funds). During QT, the effect will be greater if non-bank participants replace the Bank as the marginal buyer of GoC bonds. This is the most likely scenario since historically, non-bank participants have held a significant share of the GoC bond market. As well, during QE, the Bank primarily displaced these entities in terms of GoC holdings.

This analysis focuses exclusively on the mechanical impacts of QE and QT. Other factors that could influence the size and composition of commercial banks’ balance sheets, including natural growth in bank deposits, loan growth and changes in the level of government bond issuance, are held constant.

## The Bank’s footprint in the government bond market will shrink as quantitative tightening proceeds

The Bank held a total of about $430 billion in GoC bonds before it ended QE and entered the reinvestment phase in November 2021. During this period, the Bank kept its holdings of GoC bonds roughly constant (Chart 1). QT, launched in April 2022, began reducing GoC bond holdings on the Bank’s balance sheet by not replacing maturing GoC bonds. As a result, the size of the Bank’s balance sheet will decrease over time in a predictable manner.5 ### Chart 1: Quantitative tightening will reverse the increase in the Bank of Canada’s holdings of government bonds Source: Bank of CanadaLast observation: July 2022 While QE and QT have significant impacts on the amount of GoC bonds the Bank holds on its balance sheet, they also alter the ownership structure of the GoC bond market. The share of outstanding GoC bonds held by the Bank peaked at around 41% at the end of 2021, up sharply from 13% before the COVID‑19 pandemic (Chart 2).6 ### Chart 2: Quantitative easing increased the Bank’s share of government bonds at the expense of non-bank investors Note: Banks are domestic banks, while Non-banks include all other investors (i.e., households, pension funds, mutual funds, exchange-traded funds, foreign investors, etc.). Source: Statistics CanadaLast observation: 2021Q4 QT will reverse this trend because the Bank will no longer buy GoC securities through its QE operations and will let its existing GoC holdings mature. Around$145 billion of GoC bonds held by the Bank will mature by the end of 2023. As the Bank’s holdings of GoC bonds decline, private sector participants—such as commercial banks, households, businesses and investment funds—will need to hold more bonds7, assuming no offsetting change in the stock of GoC debt.

The historical predominance of non-bank participants in the GoC market (Chart 2) suggests that they will likely remain important buyers of GoC bonds and should see their share of holdings return closer to traditional levels.

## The mechanics of quantitative easing and quantitative tightening

This section provides some simplified examples that describe how QE and QT can affect the balance sheets of both the Bank and the Canadian banking system. These examples hold all other parts of the Bank’s balance sheet constant.

For simplicity, we start by examining the effect of QE under two extreme cases, when the Bank purchases a bond from a:

• commercial bank
• non-bank entity

### Case 1: The Bank purchases a bond from a commercial bank

To pay for the bond, the Bank deposits funds electronically into the commercial bank’s settlement account at the Bank of Canada. This creates a unique type of liability called settlement balances (or reserves).8

After the transaction, the following occurs:

• The size of the Bank’s balance sheet increases. Assets increase by the amount of bonds purchased. The purchases are funded by an equivalent amount of reserves owed to banks (a liability of the Bank) (Figure 1). Put differently, for every dollar of securities purchased, both the Bank’s assets and its liabilities increase by a dollar.9
• In contrast, the overall balance sheet of the banking system remains the same size, but the asset composition shifts. The banking system’s holdings of GoC bonds decrease while its reserves held at the Bank increase by the same amount.

#### Figure 1: The balance sheets of the Bank of Canada and the banking system change in different ways

Figure 1: The balance sheets of the Bank of Canada and the banking system change in different ways

Balance sheet activity in Case 1 (when the Bank of Canada buys a Government of Canada bond from a commercial bank)

AssetsLiabilities
Banking system
AssetsLiabilities
🡑Reserves

### Case 2: The Bank purchases a bond from a non-bank entity

In this case, the mechanics are different because non-bank participants—including households, businesses and investment funds—cannot hold settlement balances at the Bank. Two things happen simultaneously in this transaction:

• The non-bank entity sells the GoC security to a commercial bank and is credited with a deposit in its account.
• The commercial bank sells the newly acquired GoC security to the Bank and receives settlement balances in return.

In this example, the commercial bank effectively acts a conduit between the non-bank entity and the Bank.

After the transaction, the following occurs:

• Like in Case 1, the size of the Bank’s balance sheet increases by the amount of the GoC bond purchase.
• The size of the overall balance sheet of the banking system also grows because the commercial bank now holds more reserves at the Bank of Canada and owes the non-bank entity the equivalent amount in its deposit account (Figure 2).

#### Figure 2: The balance sheets of the Bank of Canada and the banking system both increase in size

Figure 2: The balance sheets of the Bank of Canada and the banking system both increase in size

Balance sheet activity in Case 2 (when the Bank of Canada buys a Government of Canada bond from a non-bank participant)

AssetsLiabilities
Non-bank participant
AssetsLiabilities
🡑Deposits
Banking system
AssetsLiabilities
🡑Reserves🡑Deposits

QT has the opposite effect. During QT, the bond holdings of the Bank will mature and roll off the Bank’s balance sheet. The government is assumed to continue to issue new bonds to fund these maturities. Overall, the size of the Bank’s balance sheet will decrease regardless of who buys these newly issued bonds. In contrast, the impact on the size of the banking system’s balance sheet will depend on who purchases these new bonds:

• If banks buy the GoC bonds, the size of the balance sheet of the banking system will remain unchanged, although the composition of assets will shift.
• If non-bank participants buy the GoC bonds, the banking system’s balance sheet will, all else being equal, decrease by the amount of GoC bonds sold. Deposits will be pulled out of the banking system to fund this transaction.10

## How quantitative easing and quantitative tightening affect the liquidity coverage ratios of banks

QE puts upward pressure on the liquidity coverage ratios (LCRs) of major Canadian banks. The LCR is a regulatory requirement that is meant to ensure banks have enough liquid assets to withstand an episode of short-term liquidity stress. To meet the LCR, banks must hold a sufficient stock of high-quality liquid assets (HQLAs), such as cash and cash equivalents, to cover net cash outflows during a 30-day liquidity-stress scenario.

$$\displaystyle LCR$$ $$\displaystyle=\,\frac{HQLA}{Net\ stressed\ cash\ outflows\ over\ \mathit{30}\ days}$$

During the initial phase of QE, the total LCR of Canada’s Big Six banks increased from 132% in the fourth quarter of 2019 to a peak of 151% in the third quarter of 2020 (Chart 3).11

### Chart 3: The Bank’s quantitative easing program put upward pressure on the liquidity coverage ratio of major Canadians banks

Note: The liquidity coverage ratio is defined as the ratio of high-quality liquid assets to total net cash outflows over the next 30 calendar days. Canada’s Big Six banks are federally regulated financial institutions that have been designated as systemically important to the Canadian financial system by the Office of the Superintendent of Financial Institutions. They are the Royal Bank of Canada, Toronto Dominion Bank, Bank of Nova Scotia, Bank of Montreal, Canadian Imperial Bank of Commerce and National Bank of Canada.
Sources: Big Six banks’ public quarterly financial statementsLast observation: 2022Q2

This growth in the total LCR was largely caused by the expansion of HQLA, which resulted from the rise in settlement balances driven by QE.12 The LCR increased because:

• In the calculation of the LCR, the full value of settlement balances qualifies as HQLA. This means that the numerator of the LCR grows as settlement balances increase.13
• In contrast, the net stressed cash outflows—the denominator of the LCR—grew by only a fraction of the total increase in bank deposits. Under the LCR, bank deposits contribute to cash outflows because some deposits are assumed to run off the balance sheet during a stress event. For example, retail deposits that are quite stable and unlikely to be withdrawn are assumed to have a run-off rate of 3% to 5%. For less stable deposits, such as non-transactional, uninsured or brokered deposits, the assumed run-off rates range from 10% to 40%.14

As a result, the net impact of QE was positive for the LCR, assuming some of the GoC bond purchases were from non-bank participants.

Over time, the banking system adjusted and the total LCR returned to levels similar to those seen before QE (Chart 3). This is because:

• Banks reduced their holdings of HQLAs other than settlement balances.
• A large fraction of the term repurchase agreements (repos) provided by the Bank matured. This reduced the amount of settlement balances in the banking system.
• Rapid loan growth contributed to a rise in deposits. Higher deposits increased the estimated value of net stressed cash outflows over a 30-day period.

Moving forward, QT will put downward pressure on the LCR. Non-bank participants will likely purchase a significant share of the new bonds issued by the government. In this case, they will draw down their deposits at commercial banks and use the money to purchase the newly issued government securities. All else being equal, the liabilities of the banking system will shrink as these deposits leave, but this will be accompanied by a simultaneous and equal reduction in assets in the form of settlement balances. This decrease in settlement balances will lower the amount of HQLA held by the banking system.

It is important to note that this simplified analysis assumes that all factors that can influence the LCR do not change and that banks are passive to the QT process—they do not change how they manage their liquidity in reaction to QT. But in reality, banks are expected to actively manage their LCRs during the QT period. For instance, they could purchase some of the newly issued GoC bonds or raise new term funding. Their strategies will be guided by the minimum regulatory requirement as well as their own appetite for liquidity risk.

## Conclusion

The operational mechanics of QE and QT affect not only the balance sheet of the Bank of Canada but also that of the overall banking system. These balance sheet effects also have an impact on the regulatory liquidity position of banks.

QE operations removed one type of HQLA (i.e., Government of Canada bonds) from the financial system and replaced it with another (i.e., overnight deposits at the Bank of Canada as settlement balances). When the Bank purchases assets from non-bank participants, this contributes to the expansion of the banking system’s balance sheet. This had a net positive impact on the total LCR of the banking system.

QT will reverse this. The impact on the banking system’s balance sheet will, all else being equal, depend on how newly issued GoC debt is allocated within the financial system. If non-bank participants increase their GoC bond holdings, deposits will be pulled from the banking system and used to purchase the newly issued government securities. Banking system liabilities will decline as deposits exit the financial system. This will also be accompanied by a simultaneous and equal reduction in settlement balances. Canadian banks that have a significant presence in the United States will be impacted through the same channels by the ongoing QT operations of the US Federal Reserve (Leonard, Martin and Potter 2017). All of these effects of QT will be felt on the liquidity position of the banking system as LCRs will experience downward pressure, which could force banks to adjust their liquidity management strategies.

1. 1. By purchasing government bonds, which are used as a reference rate for other Canadian bonds and loans, the Bank placed downward pressure on interest rates across the economy. This reduced borrowing costs and helped support economic activity and the Bank’s inflation-targeting objective throughout the pandemic.[]
2. 2. See P. Beaudry, “Our quantitative easing operations: Looking under the hood” (remarks delivered virtually to the Greater Moncton Chamber of Commerce, the Fredericton Chamber of Commerce and the Saint John Region Chamber of Commerce, December 10, 2020).[]
4. 4. This note does not address the impact of changes in central bank balance sheets on bond yields or market liquidity.[]
5. 5. See Bank of Canada, “Bank of Canada provides operational details for quantitative tightening and announces that it will continue to implement monetary policy using a floor system,” market notice (April 13, 2022).[]
6. 6. The Bank’s market share of GoC bonds averaged around 8% between March 1990 and March 2020. In normal times, the Bank purchases GoC bonds and bills in the primary market to respond to the public’s growing demand for bank notes in circulation.[]
7. 7. Otherwise, the outstanding stock of GoC debt would need to decline by an amount equivalent to the maturing bonds that were held by the Bank of Canada.[]
8. 8. As noted earlier, both Case 1 and Case 2 present stylized, simplified examples. In reality, primary dealers would act as intermediaries between the Bank of Canada and the seller.[]
9. 9. This equivalency does not hold over time because the level of settlement balances is affected by growth in other liabilities (for example, bank notes in circulation and government deposits) and changes in the level of other assets (such as term repos [repurchase agreements]).[]
10. 10. Non-bank participants could also fund a portion of these bond purchases through the bank-intermediated repo market. This would lessen the drawdown in deposits. When this occurs, the repo transaction creates a new deposit, which is then used to fund the bond purchase made by the non-bank participant.[]
11. 11. On the Canadian banking system’s balance sheet, Canada’s Big Six banks hold 93% of the overall value of assets.[]
12. 12. In addition to asset purchases, settlement balances also rose in response to the Bank of Canada’s term repo operations. See Term Repos on the Bank’s website for further details on these operations.[]
13. 13. This assumes the QE purchases were made from the non-bank sector. Purchases from the banking sector would simply replace one type of HQLA (GoC bonds) with another (settlement balances).[]
14. 14. Non-operational deposits from financial institutions have an assumed run-off rate of 100%. An asset purchase by the Bank from a non-bank financial institution has a neutral LCR impact on the banking system, assuming the proceeds of the transaction are kept as a non-operational financial institution deposit. However, during the first two years of the pandemic, deposits from financial institutions contributed to only 4% of the increase in Canadian-dollar-denominated deposits. This suggests that most of the deposits created indirectly by QE were not kept as deposits by these institutions but were returned to the banking system in the form of other types of deposits (i.e., retail, small business or non-financial corporate).[]

## References

Leonard, D., A. Martin and S. M. Potter. 2017. “How the Fed Changes the Size of Its Balance Sheet.” Federal Reserve Bank of New York Liberty Street Economics, July 2017.

Chu, P., G. Johnson, S. Kinnear, K. McGuinness and M. McNeely. 2022. “Settlement Balances Deconstructed.” Bank of Canada Staff Discussion Paper No. 2022‑13.

## Acknowledgements

We thank Russell Barnett, Toni Gravelle, Scott Kinnear, Louis Morel, Stephen Murchison, Guillaume Ouellet Leblanc, Carolyn Rogers and Jing Yang for helpful discussions and suggestions. Finally, we are grateful to Meredith Fraser-Ohman and Alison Arnot 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): E, E5, E51, G, G2, G21, G23, G3, G32

DOI: https://doi.org/10.34989/san-2022-12