How well can large banks in Canada withstand a severe economic downturn?

Introduction

Do large banks in Canada have enough capital to withstand a severe economic downturn? The COVID‑19 pandemic caused a large disruption to the Canadian economy. However, this health crisis has not turned into a financial crisis, for two main reasons. First, Canadian banks were well capitalized going into the pandemic and continued to act as a shock absorber for the economy. Second, extraordinary policy measures—fiscal, monetary and prudential—significantly reduced household and business insolvencies, softening the impact of the shock on the real economy and the financial system.

In this note, we assess whether the Canadian banking system has enough capital to withstand a severe and prolonged recession when authorities do not provide a large amount of policy support. Using our solvency stress-testing tool, we measure the effect of our risk scenario on the capital positions of major banks in Canada.1 Although this represents a hypothetical scenario and reality turned out differently, the exercise is informative in understanding potential vulnerabilities of the banking system. Importantly, we are not evaluating outcomes for individual institutions, but rather seeking to understand the resilience of the banking system as a whole.

Similar to Gaa et al. (2019), we find major banks would incur significant financial losses in such a severe scenario but would remain resilient and maintain levels of capital well above the regulatory minimum. The resilience of the Canadian banking system in this scenario can be explained by:

  • strong capital buffers and high loan-loss reserves going into the scenario, which help absorb additional loan losses. The capital buffers of Canada’s largest banks increased during the pandemic as earnings outpaced the distribution of dividends. In addition, the banks built up high reserves in the first quarters of the pandemic as a precaution, anticipating potential future loan impairments.
  • sound underwriting and risk management practices, which limit the occurrence and magnitude of defaults on bank loans to households and businesses.
  • strong initial balance sheets of households and businesses, allowing them to weather the effects of the simulated downturn.

The risk scenario

To stress test the solvency of major banks in Canada, we design a global risk scenario in which economic activity contracts sharply over an extended period. Among the possible triggers for such a severe downturn is an economic setback caused by a new wave of the pandemic.2 Our risk scenario begins with the emergence of a COVID‑19 variant spreading rapidly around the world and resistant to existing vaccines. The narrative of this scenario is realistic: the Omicron variant is a reminder that how the pandemic evolves remains uncertain.

Given the data available at the start of the stress-testing analysis (around mid-2021), the risk scenario begins in the first quarter of 2021 and lasts for three years. This scenario should not be seen as a projection of events if a new COVID‑19 variant emerges. Rather, it is hypothetical and assesses the resilience of the Canadian banking system to a severe and prolonged economic recession. Therefore, we make strong assumptions regarding:

  • lockdowns
    • Public authorities reimpose lockdowns to slow the transmission of the variant, reducing economic activity and adding uncertainty about the economic outlook.
  • policy support
    • Governments gradually unwind their support programs as previously planned and do not introduce new fiscal transfers.
    • Macroprudential policy tools, such as adjusting the domestic stability buffer, remain unchanged.
  • scarring
    • The crisis causes long-lasting consequences for the economy. For instance, unemployed workers see their skills depreciate, and they struggle to re-enter the job market. Firms that become insolvent disappear from the business landscape.

In the scenario, the lockdowns trigger a severe and persistent recession that lasts six quarters. The initial impact is limited to industries sensitive to COVID‑19 but quickly spreads to the broader economy. At the height of the crisis, Canada’s gross domestic product (GDP) contracts by 5.8% and the unemployment rate reaches a peak of 13.5% (Table 1). Also, Canadian households see their incomes decline and reduce their spending on residential investment. This results in a 29% correction of house prices at the national level.

Table 1: The risk scenario is more severe than previous recessions in Canada

Table 1: The risk scenario is more severe than previous recessions in Canada
  Decline in real gross domestic product (peak to trough, %) Duration of recession (consecutive quarters of negative growth) Increase in unemployment rate (percentage points) Peak unemployment rate (%) Decline in house prices (peak to trough, %)
Risk scenario* -5.8 6 4.7 13.5 -29.0
2008–09 recession -4.5 3 2.7 8.6 -7.8
1990–91 recession -3.4 4 4.1 11.7 -6.6
1981–82 recession -5.4 6 5.8 13.0 -9.8

* The risk scenario begins in 2021Q1. These impacts are in addition to those already experienced during the first three quarters of the COVID‑19 pandemic.
Sources: Statistics Canada and Bank of Canada calculations

About one year after the spread of the new variant, positive news on vaccines begins to surface. However, economic scarring results in a slow recovery, with GDP not returning to pre-pandemic levels throughout the three-year horizon of this risk scenario.

Relative to previous major Canadian recessions, the risk scenario is more severe, is somewhat more protracted and includes a much larger correction in house prices. The scenario is also global, with similar issues affecting economies in other countries.

The stress-testing tool

We use the Top-Down Solvency Assessment (TDSA) tool to estimate the impact of our risk scenario on the capital positions of banks.3 The TDSA can generate a projection of various components of the balance sheets and income statements of banks in Canada. As shown in Figure 1, the paths of economic and financial variables described in our scenario serve as inputs to the TDSA, which allows us to calculate key capital ratios. To assess the resilience of the Canadian banking system under the stress scenario, we then compare these capital ratios with minimum regulatory requirements.

Impact of the risk scenario on banks

The impact of the economic downturn on major Canadian banks in our scenario is evident in three ways:

  • a weakening in demand for bank credit, reflecting heightened uncertainty
  • a rise in credit losses, due to a greater number of defaults on existing loans
  • an increase in risk-weighted assets driven by a decline in credit quality

A weakening in demand for bank credit

In our risk scenario, the weaker demand for bank credit plays out differently across the various segments of their loan books (Chart 1). Banks initially experience a surge in demand for business credit as firms rely more extensively on their existing credit facilities for precautionary reasons and to compensate for a decline in their revenues (similar to what occurred early in the pandemic). Businesses eventually pay back these loans as their confidence returns. Over time, weak loan demand, due to the lengthy downturn, leads to outstanding balances falling in all the business, consumer and mortgage segments. Loan repayments and the elimination of exposures due to defaults mean that the lower outstanding balances of loans exceed new lending volumes.

Chart 1: Weak demand for credit eventually leads to lower outstanding balances across all loan categories

Source: Bank of Canada calculations

A rise in credit losses

The declines in employment and economic activity in the risk scenario create difficulties for some borrowers to repay their bank loans, causing some of them to default on their debt. Banks are required to set aside funds to deal with not only these defaulting borrowers, but also borrowers that are expected to default in the future.4 Collectively, these funds are called provisions for credit losses (or simply credit losses), which tend to increase as the risks increase for banks’ assets. We estimate that banks would face credit losses totalling $123 billion over the three-year scenario. When expressed as a percentage of initial loan balances, this dollar amount represents a credit loss rate of 4.4%. This is higher than loss rates experienced during previous economic downturns in Canada (Chart 2).5

Chart 2: Estimated credit loss rates in the risk scenario are higher than observed during previous major downturns

* Credit losses were realized between 1993Q3 and 1994Q2.
Sources: Regulatory filings of Canadian banks and Bank of Canada calculations

Our results show important differences in three-year credit loss rates across the various product lines offered by banks, such as:

  • credit cards. The credit loss rate assumed for this product line is the highest, at 35%. This reflects a higher probability of default on credit card balances and minimal recovery by banks in the event of a default.
  • uninsured mortgages. We estimate that the credit loss rate is relatively small, at 0.7%.6, 7 This reflects mainly:
    • sound underwriting standards leading to low default rates. For example, when issuing an uninsured mortgage, banks must ensure that households can service their mortgage (e.g., borrowers must pass a mortgage stress test under the minimum qualifying rate) rather than relying solely on the value of the underlying collateral for protection.
    • ample home equity for uninsured mortgages (typically at least 20%) that absorbs much of the decline in house prices and cushions banks from incurring large losses in the event of a default.8
  • business loans. We assume that the credit loss rate is somewhat higher than historical experience. Although these losses are experienced differently across sectors, real estate sectors (e.g., construction) and those sensitive to COVID‑19 bear an outsized share of these losses in our scenario.9

An increase in risk-weighted assets

In the regulatory framework of banks, the concept of risk-weighted assets (RWAs) is used as the denominator to calculate capital ratios. In a downturn, the risk profile of existing loans deteriorates, shifting the composition of bank assets toward higher risk weights in the calculation of RWAs.10 In our risk scenario, RWAs grow by a total of 22%, primarily driven by the higher risk weights applied to existing loans. However, this growth is partly offset by lower loan balances.

An increase in the value of RWAs implies that, all else being equal, banks must hold more capital to meet regulatory requirements.

Impact on the regulatory capital ratios of banks

In Canada, banks are required to maintain a level of capital on their balance sheets above certain regulatory thresholds. In particular, banks must hold enough common equity Tier 1 (CET1) capital to represent at least 4.5% of their risk-weighted assets.11 They must also keep additional capital buffers, such as:

  • a capital conservation buffer (CCB) of 2.5% of RWAs. The CCB allows banks to absorb losses by limiting capital distributions in the event of a breach. Breaching the CCB would result in automatic restrictions on the size of dividends to shareholders and share buybacks. In practice, given the long-standing tradition of dividend payments by major banks, they would likely take some actions to support their capital positions before breaching this buffer.12 However, in our scenario analysis, we assume that banks do not take such actions.
  • an additional CCB surcharge of 1% of RWAs. This is because the major Canadian banks included in our stress-testing exercise are designated as domestic systemically important banks (D-SIBs).13
  • a domestic stability buffer (DSB), which ranges between 0% and 2.5% of RWAs. The DSB varies depending on the health of the economy, requiring banks to maintain higher buffers of capital when times are good. When times are not as good, the Office of the Superintendent of Financial Institutions can reduce the required buffer of capital, freeing up additional lending capacity for banks to support the recovery. In our scenario, the DSB is kept constant at 1%, which was the level in the fourth quarter of 2020.

In the risk scenario, capital impacts are significant, with banks breaching the CCB requirement of 8% during the first two years. The aggregate CET1 capital ratio declines rapidly from 12.3% to 7.4% in the first year (-4.9 percentage points) before recovering to 8.3% by the end of the scenario (Chart 3). Although the decline in capital ratios is material, when considering the size of the economic shock, banks are viewed as remaining broadly resilient.

Chart 3: In the risk scenario, capital ratios breach the capital conservation buffer but remain above the regulatory minimum

* The yellow horizontal line reflects the domestic stability buffer level that was effective in 2020Q4 and the level assumed in our risk scenario.
† The capital conservation buffer includes the additional surcharge for domestic systemically important banks.
Source: Bank of Canada calculations

As a result of their pandemic response, banks enter the risk scenario with a stronger capital position than they had before the pandemic, which contributes to the resilience of Canadian banks in our scenario.14 Their aggregate CET1 capital ratio increased from 11.6% in the first quarter of 2020 to 12.3% in the fourth quarter of 2020. This represents roughly $15 billion in additional capital. Moreover, banks set aside $13 billion in additional precautionary provisions for credit losses throughout 2020, resulting in a total of $24 billion in loan loss reserves that help offset losses in our stress test starting in the first quarter of 2021.

Chart 4 shows the decline in the aggregate CET1 capital ratio between the initial conditions (in the fourth quarter of 2020) and the end of the risk scenario (in the fourth quarter of 2023). The strong capacity of banks to generate internal revenues, or pre-provision net revenue, adds 7.7 percentage points to the CET1 capital ratio over the scenario. Banks’ capacity to maintain robust internal revenues despite the severe downturn can be explained by two facts:

  • Canadian banks start from a position of high profitability, typically posting returns on equity greater than 15% in normal times.
  • Canadian banks operate under a highly diversified model with significant retail and commercial lending operations, capital markets operations, wealth management arms and insurance businesses. This diversification helps ease the impacts on earnings when a particular business line is struggling.

However, while these operating revenues help banks absorb some of their losses in the scenario, they are roughly one-third less than what they would be in normal times. This is because of:

  • a compression of their net interest margin given the continued environment of low interest rates
  • weaker lending activity resulting in fewer loans to earn revenue from
  • weaker client activity affecting non-interest income

Chart 4: The decline in the common equity Tier 1 capital ratio in the risk scenario is primarily due to higher credit losses

Source: Bank of Canada calculations

This positive contribution to the CET1 capital ratio from net revenues is more than offset by:15

  • a rise in credit losses. The provisions for credit losses on performing and non-performing loans add up to $123 billion in the scenario. This contributes to reducing the aggregate CET1 capital ratio by 6.8 percentage points.
  • a steady level of dividend payments. In the scenario, we assume banks maintain the payment of dividends at levels reported in the fourth quarter of 2020, with the exception of automatic restrictions on dividends imposed because of a breach of the CCB during seven quarters. This results in $67 billion in dividends paid to shareholders throughout the three-year scenario, thereby lowering the aggregate CET1 capital ratio by 2.9 percentage points.
  • an increase in risk-weighted assets. As mentioned, RWAs increase due to the weakening of credit quality across portfolios. This leads to a decline in the aggregate CET1 capital ratio of 1.9 percentage points.

Conclusion

In this note, we stress test large banks in Canada to assess their resilience. We find that overall, major banks are strong enough to withstand a severe economic downturn. Their resilience is supported by their solid initial capital positions, a robust capacity to generate revenues even in times of stress and their reliance on sound underwriting practices.

Stress-testing frameworks are a valuable tool used by authorities to assess the capital adequacy of banks. Importantly, our stress-testing exercise provides insights into the ability of Canadian banks to support the economy during challenging times by continuing to lend to households and businesses. We note that the risk scenario for this exercise is deliberately severe and assume no reaction on the part of banks to protect their capital positions. In reality, banks themselves would likely adopt various strategies to reduce the impact of the economic downturn on their capital positions. This reinforces the conclusion that the banking system in Canada appears resilient to a large adverse shock. The resilience of Canadian banks is key to preserving the stability of the financial system.

  1. 1. In Canada, the six largest banks play a dominant role in financial intermediation. Together, they hold over 90% of the total assets of Canadian deposit-taking institutions. They are the Bank of Montreal, the Bank of Nova Scotia, the Canadian Imperial Bank of Commerce, the National Bank of Canada, the Royal Bank of Canada and the Toronto-Dominion Bank.[]
  2. 2. The pandemic-driven narrative is worked into results both through the initial calibration of macroeconomic variables and, more directly, by additional judgment about the loss rates of loans in industries sensitive to disruptions from COVID‑19.[]
  3. 3. A forthcoming technical report will provide details of the Bank’s TDSA. See also MacDonald and Traclet (2018) for a brief overview of the tool.[]
  4. 4. Under International Financial Reporting Standard 9, introduced in 2018, the accounting treatment of credit losses is more forward-looking than before. Banks must set aside funds to cover expected credit losses on all loans before those losses are incurred. This allows the banks to recognize losses earlier than they were able to under the previous accounting standard.[]
  5. 5. In this exercise, we consider credit losses associated with loans, including losses from committed lines of credit. We also account for credit risk associated with securities exposures under market losses; however, we do not consider either counterparty credit risk associated with derivatives or securities financing transactions.[]
  6. 6. Note that more than 40% of mortgages held by major banks are protected by mortgage insurance, implying that the insurers bear potential losses on these mortgages rather than banks.[]
  7. 7. Estimated loss rates on Canadian mortgage exposures, although consistent with historical experience, are potentially low given the lack of large, observable financial stress leading to major losses on mortgages in Canada. As such, we perform additional sensitivity analysis to ensure that results hold, even for larger loss rates. As presented, our work assumes a 21% loss on defaulted uninsured residential mortgage exposures, which is calibrated based on the fixed costs for recovering the value of the home (e.g., sales cost) and the expected negative equity position at the time of sale.[]
  8. 8. Canadians build up significant equity in their homes over the years due to a minimum requirement of at least a 20% down payment on a property financed with an uninsured mortgage, typical amortization schedules with a gradual repayment of the mortgage over time and increases in house prices.[]
  9. 9. See Bruneau, Duprey and Hipp (forthcoming) for more information on how the corporate probability of defaults are projected in our solvency stress test. Corporate probability of defaults is a key input to estimate credit losses on business exposures.[]
  10. 10. In a period of economic stress, the increase in risk weights is particularly noticeable in loan categories that traditionally carry lower credit risk, such as residential mortgages.[]
  11. 11. CET1 largely consists of ordinary shares, retained earnings and accumulated other comprehensive income.[]
  12. 12. Actions taken to support capital positions could include reducing expense spending, adjusting pricing to limit balance sheet growth, making use of dividend reinvestment plans, issuing new shares and selling non-core assets.[]
  13. 13. A D-SIB could disrupt the functioning of the domestic financial system should it face solvency issues.[]
  14. 14. The stronger starting position for capital and loan loss reserves is a result of the pandemic response of banks rather than an attempt to gain an improved starting point in our stress scenario.[]
  15. 15. Our simulation shows the first-round effects through credit, income and market losses. As capital ratios remain above the regulatory minimum throughout the scenario, second-round effects, such as fire-sale losses from forced deleveraging or losses on interbank exposures from contagion effects, do not apply and are therefore not shown here. Second-round effects are estimated using the Bank of Canada’s MacroFinancial Risk Assessment Framework (see Fique 2017).[]

References

Bruneau, G., T. Duprey and R. Hipp. Forthcoming. “Forecasting Banks’ Corporate Loan Losses Under Stress: A New Corporate Default Model.” Bank of Canada Technical Report.

Gaa, C., X. Liu, C. MacDonald and X. Shen. 2019. “Assessing the Resilience of the Canadian Banking System.” Bank of Canada Staff Analytical Note No. 2019-16.

Fique, J. 2017. “The MacroFinancial Risk Assessment Framework (MFRAF), Version 2.0.” Bank of Canada Technical Report No. 111.

MacDonald, C. and V. Traclet. 2018. “The Framework for Risk Identification and Assessment.” Bank of Canada Technical Report No. 113.

Acknowledgments

We thank Russell Barnett, Thibaut Duprey, Toni Gravelle, Grzegorz Halaj, Louis Morel, Stephen Murchison, Carolyn Rogers, Tamara VanDeWalle and Jing Yang for helpful comments and suggestions. Finally, we are grateful to Alison Arnot and Maren Hansen for editorial assistance and Gabriel Bruneau and Sofia Priazhkina for technical 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, E2, E27, E3, E37, E4, E44, G, G1, G2, G21, G23

DOI: https://doi.org/10.34989/san-2022-6