The pandemic-induced weakening asset quality among the five largest Brazilian banks in 2020 prompted them to increase substantially provisions for credit losses, jeopardizing bottom-line results. Our base-case scenario assumes that banks' asset quality will continue hurting profitability in 2021 but to a lesser extent than in 2020. Relaxed social-distancing measures and the start of vaccinations should help the country's economy start to recover after the severe recession last year.

On the other hand, 2021 presents a high degree of uncertainty, given that the second virus wave that Brazil is facing may require the reimposition of economically painful lockdowns, while the government has limited fiscal capacity to help the ailing economy. With these factors in mind, we have established three stress scenarios to estimate potential losses and to what extent the large banks' capital adequacy will be squeezed.

Credit Growth Picked Up In 2020

Lending growth in Brazil was strong during 2020, mainly because of the higher demand from corporations intent on cushioning their liquidity to weather the economic shock. In addition, the central bank provided incentives to boost lending through liquidity support, loan guarantees, and capital and provisioning relief. As a result, we expect a 14% year-over-year credit growth last year, which will be the first time large banks have posted double-digit growth on aggregate since 2015. However, we expect lending growth to ease in 2021 to about half of last year's pace given the absence of the extraordinary government measures and due to banks' reluctance to take on additional risk, given the potential second wave of infection and economic uncertainty it's generating. Moreover, we expect retail unsecured lending to continue posting modest growth as supply and demand for some lending products, such as credit cards and auto loans, are still lower than prior to the pandemic, while demand for residential mortgages and payroll deductible loans continue booming.

Chart 1

Provisioning Approaching The Historical Peak

Since 2017, Brazilian banks had been recovering from the severe economic downturn between late 2014 and 2016, which caused provisions for bad loans to spike and bottom-line results to fall. Prior to the pandemic, loan-loss reserves dropped to nearly a half of the amount provisioned at the depth of Brazil's biggest recession in the century. Lower reserves was a result of improving conditions in the corporate sector, banks' tighter loan underwriting standards, and shift away from riskier borrowers. Portfolio mix also has gradually shifted to less risky products, such as payroll loans and residential mortgages. In addition, banks increased the use of loan guarantees to corporations, and small to midsize enterprises. However, in the first half of 2020, combined loan-loss reserves among large banks were about 50% higher than in a year-earlier period amid the deteriorating credit conditions in Brazil, because of the COVID-19 pandemic and the pernicious economic impact from containment measures. These figures were very close to the September 2015 peak.

We believe government policies were instrumental in easing the effects of higher credit risk, while the central bank has taken a timely and comprehensive set of measures to mitigate the harsh impact of COVID-19 on banks' asset quality, which diminished provisions in the second half of the year. In addition, the government has introduced extraordinary support to relieve pressure on struggling households. The fiscal support package accounts for 12% of the country's GDP, one of the largest among emerging markets. As a result, we expect large Brazilian banks' credit losses to remain manageable thanks to conservative growth strategies in the past two years and the regulatory measures that help cushion the economic damage from the pandemic.

Chart 2

Stress Test Scenarios

Based on these factors, we have established three stress scenarios to estimate potential losses and strains on large banks' capital adequacy. Each scenario has two variables: loan-loss reserves coverage and nonperforming assets (NPAs) as follows:

• Stress C: We're keeping loan-loss reserves coverage at the pre-pandemic level and raising the 2019 NPAs level by 25%.
• Stress B: We're keeping the loan-loss reserves coverage at the 20-year average and lifting the average NPAs level for the last 20 years by 50%;
• Stress A: We're increasing the highest historical NPAs level for the last 20 years by 50% and keeping the loan-loss reserves coverage at 1.0x.

The Potential Contraction In Regulatory Capital Ratios Is Manageable

As a result of weaker profitability and rapid credit growth, large banks' regulatory capital and Tier 1 ratios are likely to decrease 40-80 basis points (bps) in 2020 from the 2019 levels. Our base-case scenario assumes that the 2021 capital ratios will be stable because of lower lending growth and modest credit provisioning. However, uncertainties for performance in 2021 exist, given the second wave that Brazil is currently facing that could impair economic activity.

The following are the results of our stress test for 2021, including three scenarios:

• In stress C scenario, we project regulatory capital ratios to fall 150 bps. We believe this scenario would be feasible under relaxed social-distancing measures that could last the entire year, given that the efficacy of the vaccines that Brazil has just started to validate is still uncertain. This scenario also incorporates the continuation of government extraordinary support in 2021, which would mitigate the pandemic's economic and social harm.
• In Stress A and B scenarios, we project regulatory capital ratios to erode 250-300 bps. We believe these scenarios would be feasible under more severe social-distancing measures along with quarantine and lockdowns. We considered that under these scenarios, the government would offer limited or no extraordinary support, which would exacerbate the pandemic's economic and social effect.

Chart 3

Chart 4

*For this exercise we considered the Basel III and Tier 1 capital requirements of 10.25% and 8.25% respectively since the central bank has temporarily reduced the conservation buffer requirements from 2.5% to 1.25%.

The drop in regulatory ratios could be steeper or milder than the average if we consider each bank individually based on its historical credit losses, coverage level, and capital position. In the chart, we have the disclaimer of the "highest" and "lowest" indicator in each scenario. Based on our analysis, we only had one bank breaching the minimum Basel III ratio under stress A scenario. Typically:

• Banks that had an outlier peak in their NPL historical data would have lower regulatory capital ratios under stress A scenario;
• Banks that had consistently high NPLs in their average historical data would show a weaker result under stress B scenario; and
• Banks that had weaker-than-average coverage on their balance sheets prior to the pandemic would have a weaker result under stress C scenario.

Ratings Should Move In Tandem With Those On The Sovereign Despite Potentially Lower RAC Ratios

To evaluate the banks' capitalization level, we apply our risk-adjusted capital framework (RAC) globally, regardless of regional regulations and banks' internal risk measures. Our RAC ratio compares our definition of total adjusted capital to our risk-weighted assets, reflecting a risk metric that's globally more comparable than regulatory ratios. The main difference regarding our methodology and local regulation is the higher risk weights we apply to deferred tax assets of temporary differences from what the regulator applies; and the high risk weights we apply to the sovereign exposures based on our sovereign rating, while the regulator applies no risk weight.

Our base-case scenario assumes that Brazilian large banks' RAC ratio will remain moderate, according to our methodology in the next two years. This implies an average RAC in the 5%-7% range. Under our stress test scenarios, the RAC would erode at a similar pace as the regulatory capital ratios. Under all three stress scenarios, large banks' RAC ratios would drop from the moderate threshold to the weak one, according to our methodology, which we define at the 3%-5% range. Under stress C scenario, however, the RAC ratios would be very near the 5% threshold, while under stress A and B scenarios, the metrics would be closer to 4%.

Chart 5

Given potentially weaker capital ratios under each stress scenario, we expect a bank's stand-alone credit profile (SACP) to drop one notch under stress C scenario, and a two-notch drop under stress A and B scenarios. Nevertheless, despite potentially weaker stand-alone financial profiles, the higher SACPs would cushion ratings on large banks. Therefore, we would expect our ratings on these banks to continue to reflect that on Brazil under all stress scenarios. The ratings on the sovereign constrain those on the banks, given their exposure to the domestic market. Therefore, we would still expect ratings on banks to move in tandem with the sovereign ratings.

Large Banks Are Prepared To Face Economic Fallout From The Second Wave

Large Brazilian banks' profitability took a big hit in 2020 due to heavy provisioning. We believe that a sustained drop in economic activity due to long-running lockdowns and restrictions could increase the volume of problematic loans sharply and continue to hurt banks' asset quality in 2021. However, despite a potentially considerable drop in regulatory capital ratios, we believe that large Brazilian banks can weather the economic damage from the second pandemic wave. We estimate that the buffer these banks have to cope with credit losses before reaching the minimum level of capital requirements is about R$280 billion. This is a result of high profitability and sound capital adequacy, which allowed large banks to pass our hypothetical stress scenarios without any regulatory breach.

Moreover, we believe large private banks have the capacity to reduce their cost structures and seek greater fees and commissions to maintain revenue base in a stressed economic conditions. Large public banks, on the other hand, will likely continue focusing on their core businesses and could benefit from asset sales that were postponed last year due to volatile capital market conditions. As a result, we believe that large banks' profitability will continue to support capital metrics in 2021.