Consecutive interest rate rises in quick succession have caused a large uptick in mortgage repayments for many borrowers. This is happening amid greater cost-of-living pressures, constraining household budgets. The Australian residential mortgage-backed securities (RMBS) sector has been at the forefront of these challenges, given its large exposure to variable-rate loans. This is particularly acute for more recent vintages where underlying loans were underwritten in a period of historically low interest rates.

While we believe the cash rate is nearing its peak, mortgage arrears are a lagging indicator and will continue to rise as the cumulative impact of multiple rate rises emerges. To understand how resilient RMBS ratings are to rising arrears, we stressed a representative sample of rated prime and nonconforming Australian RMBS transactions from recent vintages under scenarios with rising arrears. Our hypothetical scenarios included increased levels of loans that are 90+-days in arrears.

Scenario Details

The Reserve Bank of Australia cut interest rates to historically low levels with the onset of the pandemic in early 2020. The central bank maintained these levels until May 2022 when it lifted the cash rate, kicking off a monetary policy tightening cycle that is still underway.

Loans underwritten during this period appear to be more exposed to rising arrears, given the large increase in mortgage rates over a short period. These loans are more highly represented in recent transaction vintages. We have selected a sample of prime and nonconforming transactions across these vintages where most loans were underwritten during the low interest rate period between March 2020 and April 2022.

Across this sample of transactions, the average exposure to fixed-rate loans is minimal at about 6% for prime and 0.2% for nonconforming transactions.

For this scenario analysis, to simulate the impact of increased loans that are 90+ days in arrears, we stressed a representative sample of prime and nonconforming transactions by increasing the ratio of such loans by 50, 100, and 150 basis points (bps). We also ran an additional stress scenario for the nonconforming transactions in the sample by increasing the ratio of loans 90+ days in arrears by 300 bps.

Scenario Analysis Results

Generally, the results of these tests indicate that, even under increased economic stress, most rated tranches demonstrate ratings stability (see tables 2 and 3).

Assumptions For Scenario Analysis

For this scenario analysis, we focused on a representative sample of 38 transactions featuring prime and nonconforming collateral. The sample included transactions from recent vintages including 2021 and 2022 vintages where underlying loans were underwritten during a period of low interest rates.

We performed our scenario credit analysis in line with our criteria by applying a default frequency of 100% in all rating scenarios for borrowers that are 90+ days in arrears (see "Australian RMBS Rating Methodology And Assumptions", published Sept. 1, 2011 ). So, for example, a stress test involving a 150 bps increase in loans in 90+ days arrears resulted in a 150 bps rise in the assumed foreclosure frequency for all our rating scenarios.

Our analysis is based on loan data as of February 2023. We compared credit support requirements for the ratings on 274 tranches in the representative sample for each scenario against credit support provided by subordination. We separately analyzed cash flow for a select number of transactions where hard credit support was insufficient under the stressed scenarios to determine whether excess spread, in addition to note subordination, was sufficient to maintain current ratings on tranches under the various stress scenarios. We modeled cash flow in accordance with our cashflow modeling methodology (see "Methodology And Assumptions For Analyzing The Cash Flow And Payment Structures Of Australian and New Zealand RMBS," published June 2, 2010).

Table 1

Increase in 90+ days in arrears loans (scenarios 1-4)

For the increase in 90+ days in arrears loans scenarios, we stressed our RMBS ratings by increasing the final weighted-average foreclosure frequencies (WAFFs) by 50, 100, and 150 bps for scenarios 1-3, respectively, at each rating level. These scenarios would therefore lead to more required credit coverage at each rating level.

For example, if the 'AAA' WAFF (i.e., the gross default rate for the RMBS portfolio over the transaction's life in a 'AAA' scenario) was 15%, it increases to 16.5%% under scenario 3. We ran an additional scenario, scenario 4, for the nonconforming transactions, increasing the final WAFF by 3 percentage points at each rating level.

Some limitations and caveats to our scenario analysis

• The stresses we selected for each scenario are hypothetical and are not meant to be predictive or part of any outlook statement.
• In our scenario analysis, we applied increases in arrears immediately, which is a conservative assumption. An actual future build-up in arrears is likely to occur gradually, during which the RMBS tranches are likely to build up credit enhancement due to deleveraging, potentially lessening the ratings impact.
• The stresses are also not meant to calibrate to any of the economic scenarios we associate with our ratings in our ratings definitions.
• The results are based on the application of the models we use to rate Australian RMBS transactions. A rating committee applying the full breadth of S&P Global Ratings' criteria and including qualitative factors might, in certain instances, assign a different rating than the quantitative analysis alone would indicate.

Scenario Analysis Findings

Prime results

Table 2

Nonconforming results

Table 3

The results shown in tables 2 and 3 only consider credit support provided by note subordination and do not give any credit to available excess spread in the sample transactions. As most of these transactions were issued recently, the build-up of credit support is not as high as more seasoned transactions. We undertook cash flow modeling on 7 transactions that were failing multiple scenarios based on note subordination only. After modeling the cash flow of these transactions, our results showed that most of these tranches retained the current ratings on them under the various stress scenarios outlined above. We expect that given the excess spread available in most transactions, and structural features to enhance the utilization of excess spread in these deals, most ratings in this sample would not be affected by higher default frequencies applied under the stress scenarios outlined in our cash flow modeling analysis.

In a small number of instances, yield strain in some nonconforming transactions is adding to ratings sensitivity by limiting the excess spread available to cover higher losses. Under our stress scenarios, arrears increases are applied immediately. In practice, arrears build up gradually during which time RMBS tranches are likely to build up credit enhancement through deleveraging. As transactions in this sample are recently issued, credit support build-up is not significant for lower rated tranches of notes. As such, we think that in practice, the ratings impact will lessen over time as credit support builds, offsetting increasing arrears. However, ratings on a small number of nonconforming tranches are more susceptible to downgrades, given yield pressures that may be exacerbated if future collateral performance is worse than initial rating expectations.

Factors contributing to ratings stability:

Excess spread levels.  Excess spread levels are robust for most transactions with weighted average interest rates on underlying loans in the sample transactions ranging from 5.39%-6.93% for prime and 6.78%-7.49% for nonconforming transactions.

Structural robustness.  Many transactions, particularly nonconforming transactions, benefit from structural features that help preserve and efficiently utilize available excess spread. These include reverse turbo features (enhanced ratings stability for lower-rated notes), amortization mechanisms, and loss reserves.

Elevated prepayment rates.  High prepayment rates in this sample which range from 17%-47% across the prime sample and 28%-50% (based on February 2023 data) across the nonconforming sample, are enhancing credit support levels despite the transactions being recent.

Enhanced credit support for senior notes.  Many issuers provide extra credit support for senior notes above credit enhancement requirements. This is a common market practice in Australia.

Modest loan-to-value (LTV) profile.  Exposure to high LTV loans in the sample transactions is not significant (see chart 3).

Low exposure to fixed-rate loans.  This reduces the risk of payment shock as borrowers on variable rates have had to adjust to regular monthly increases in their mortgage repayments over the past 10 months.

Historically low unemployment rate.  Low unemployment means borrowers facing financial pressures have more options to alleviate financial stress by, for example, taking on a second job.

Australian RMBS Arrears History

The prime Australian RMBS sector has a solid track record of strong collateral performance, demonstrated by its long history of low arrears and losses. This partly reflects the relatively robust lending standards across most originators in the Australian mortgage market, and the mostly benign economic conditions in the country over the past 20 years.

Prime RMBS arrears peaked in Australia in the period immediately following the 2007-2008 global financial crisis (chart 1). Prime loans underlying Australian RMBS transactions more than 30 days in arrears peaked at 1.84% in January 2009 and loans more than 90 days in arrears peaked at 0.80% in February 2009 (chart 1). This followed a period of rising interest rates and a 1.3 percentage point increase in the unemployment rate over the prior 12 months. Prime RMBS arrears and interest rates are closely correlated given the higher variable-rate exposure across the Australian RMBS sector.

Chart 1

The arrears history of nonconforming loans in the Australian RMBS sector is more volatile. This reflects the much smaller segment of loans resulting in greater volatility in arrears levels in percentage terms. Nonconforming loans peaked at 17.09% in January 2009. Loans more than 90 days in arrears peaked at 10.54% in March 2009 (see chart 2).

We have considered historical performance in selecting the scenarios, and consider scenarios 3 and 4 to be consistent with an economic environment with materially higher levels of unemployment than the prevailing conditions and higher than S&P Global Ratings' forecasts for 2023 and 2024.

Chart 2

We do not expect nonconforming arrears to reach these peaks in the current economic conditions for the following reasons:

• Lending standards, and in particular, documentation standards have improved since the financial crisis, following the introduction of the National Credit Code.
• Total nonconforming loans outstanding were about A$4 billion in January 2009, versus A$16 billion as of March 31 2023. A small universe of loans will magnify arrears increases in percentage terms due to the smaller denominator.
• Unemployment was higher during this period (i.e. about 6% versus 3.7% currently).
• Exposure to high LTV loans has declined since the financial crisis.

Advances in technology have also enabled lenders to be more proactive with arrears management. Greater utilization of technology in enabled lenders to increase the frequency of communication with borrowers and has facilitated scalability. This can help prevent loans advancing to more severe arrears categories, and enable resources to be diverted to the more labor-intensive process of addressing advanced arrears stages.

Losses have remained low despite upticks in arrears

Since the financial crisis, cumulative gross losses (i.e. losses after the sale of the property but before excess spread or lenders' mortgage insurance) have remained low across most transaction vintages. This in part reflects the general reduction in higher LTV lending following the financial crisis, reducing the magnitude of losses in the event of borrower default. According to S&P Global Ratings pool statistics, approximately 37% of nonconforming loans and 19% of prime RMBS loans had LTV ratios greater than 80% in 2009, compared with 9% and 7%, respectively, as of March 31, 2023.

S&P Global Ratings forecasts Australia's unemployment rate will average 4.0% over 2023, rising to an average of 4.3% during 2024. While we expect unemployment to increase as the economy slows, our forecast is still below pre-pandemic unemployment levels. Based on our unemployment rate forecasts, we do not expect defaults to materially increase over the next 12-18 months.

Property price declines pose less risk as immigration resumes

According to CoreLogic's May national Home Value Index (HVI), dwelling values have declined 6.8 % over the past 12 months. Across the sample of transactions, the exposure to high LTV loans (i.e. loans with an LTV greater than 80%) is not significant (see chart 3).

Chart 3

Under our RMBS methodology, we apply a market-value decline stress to all loans ranging from 30% at 'B' rating stress up to 45% at a 'AAA' rating stress. Property price declines to date have been well within these stresses. While rising interest rates and slowing lending volumes continue to suppress demand for property, the resumption of immigration and supply shortfalls are causing property prices to rise as reflected in the 1% quarterly increase in national home values as at May 2023, according to CoreLogic data. As such, we believe the risk to our selected transactions (and the broader RMBS sector), of property price declines that could cause losses in the event of borrower defaults is low.

Interest rates, inflation, and the path less traveled

The economic path from here on is uncertain. Inflation may have peaked, but it is a long way from the Reserve Bank of Australia's target range of 2%-3%, and could take some time to return to "normal". This means interest rates may remain at higher levels for some time yet, particularly if the economy continues to demonstrate resilience on the employment front.

Despite rising arrears, RMBS ratings have been robust in the face of these challenges, thanks to low unemployment, competitive refinancing conditions, credit support levels and RMBS' strong starting position going into this monetary policy tightening cycle, aided by generous savings buffers for many borrowers. The positive rating bias across the sector over the last 18 months supports this assessment. While the road ahead is uncertain, ratings will likely remain stable for most tranches, given the credit support available and structural agility inherent in many transactions. The stalling in property price declines will also enhance borrower optionality in the event of financial stress, helping to limit losses in the event of borrower default.

Related Criteria

• Australian RMBS Rating Methodology And Assumptions, Sept. 1, 2011.
• Methodology And Assumptions For Analyzing The Cash Flow And Payment Structures Of Australian and New Zealand RMBS, June 2, 2010

Related Research

• Why Servicer Transition Risk Is Low In The Australian Structured Finance Sector, April 30, 2023
• How Much Shock Can U.K. RMBS Take?, Mar. 1 2023
• An Overview of Australia's Housing Market and Residential Mortgage-Backed Securities, Nov. 27, 2022
• RMBS Performance Watch: Australia, published quarterly
• RMBS Arrears Statistics: Australia, published monthly