The coronavirus pandemic has had an acute impact on the British economy. The Office for National Statistics reported in early August that GDP dropped by a steep 20.4% in the second quarter, while the Markit/Halifax, Nationwide, and LSL Property Services/Acadata indices recorded a negative monthly price variation in May.

The British government and the Financial Conduct Authority (FCA) rolled out several measures to prevent a sudden, sharp rise in unemployment and mortgage arrears. The Coronavirus Job Retention Scheme (CJRS), also known as the furlough scheme for employees, and the Self-Employment Income Support Scheme (SEISS), also known as the furlough scheme for self-employed workers, provide assistance for those facing furlough and reduced incomes, while the government's guidance on payment holidays addresses the impact on the residential property market. As of May 28, 2020, 17% of mortgage borrowers in the U.K. were on payment holidays, according to U.K. Finance, the trade body. By residential mortgage-backed securities (RMBS) transaction, payment holiday utilization to date shows significant variation, ranging from around 5% to 50%.

U.K. Payment Holiday Peak Has Passed

Borrowers in financial difficulties related to the pandemic may apply for a payment holiday until Oct. 31, 2020, for a period of up to three months plus a three-month extension. During this time, mortgage payments may be partially or entirely suspended, but interest will still accrue resulting in higher repayments later.

Among our rated U.K. RMBS sample, the proportion of borrowers on payment holidays is roughly in line with the overall mortgage market, though it varies widely across loan vintages and borrower type. We calculated that 16.7% of borrowers were on payment holidays on average, but the proportion ranges from 5% to more than 30% by vintage (see charts 1 and chart 2). This likely reflects underwriting standards and the intense mortgage lending competition in recent years coupled with stretched affordability. In addition, for the pre-crisis vintages, some of these borrowers are mortgage prisoners (see "More Than One-Third Of U.K. Legacy Borrowers Are 'Mortgage Prisoners'," published Sept. 4, 2019).

The Situation Is Stabilizing

We believe that the U.K. is past the peak of payment holiday utilization. However, there is a time lag between repayment of payment holidays and the latest data available for loan pools. Chart 1 shows indicative ranges for borrowers who have resumed full payment of their monthly contractual installment following the end of their initial three-month payment holiday. Borrowers who have not resumed payment would typically either have extended their payment holiday for a further three months or bounced their direct debit. For borrowers who have bounced, it is likely that once they are contacted by their servicer, they will be granted a retrospective three-month payment holiday. Therefore, in the time being, we do not expect these borrowers to be classified as in arrears.

Chart 1

Servicers have cautioned that it is possible borrowers who took a payment holiday earlier in the lockdown may have had less need than borrowers who took one later. This is because the details of the employee support scheme were not finalized then, and the downside of taking a payment holiday (capitalization and a larger ongoing payment) were not necessarily apparent to borrowers. Although at present there is no reason to believe the rate of borrowers who resumed payments in August would be materially lower than in previous months, it is possible that rates may not be as favorable (see chart 3).

Chart 2a

Chart 2b

Chart 3

Chart 3 shows that only a small proportion (1.7%) of all loans are both on payment holidays and in arrears. However, out of all the loans on payment holidays (16.7%), the aforementioned proportion accounts for more than 10% of this percentage. We have previously observed that a number of nonconforming borrowers are struggling to keep current with their payments. This indicates that there was stretched affordability prior to COVID-19, and it is likely that transactions with larger portions of borrowers with recent arrears will see higher delinquency increases than others (see "U.K. Nonconforming RMBS: One In Four Borrowers Are Just About Managing To Make Their Monthly Payments," published June 3, 2019). Lenders granted payment holidays to borrowers in arrears when they met the requirements in the FCA's guidelines. During this period, the lenders did not add the deferred payment to their existing arrears balance.

Nonconforming Owner-Occupied Borrowers More Likely To Be On Payment Holidays

We performed a logistic regression on our universe of 369,000 residential loans that report payment holidays at the loan level (see the appendix for details on our methodology). We determined that five variables primarily indicate whether a borrower is likely to be on payment holiday: current loan-to-value (LTV), debt-to-income (DTI), reperforming arrangements (including performance arrangements incurred by borrowers over the last three years or restructuring arrangements over the lifetime of a loan), previous arrears, and current interest rate. These variables may also help differentiate whether a borrower is nonconforming or prime.

Chart 4 shows the value of these variables for owner-occupied loans. For each variable, loans on payment holidays show a higher value, reflecting lower credit quality.

Chart 4

Loan vintage has a clear effect on the current LTV as most secured property values significantly rose over the last 10 years. However, the distribution shows that owner-occupied loans on payment holidays have higher current LTVs, a sign of greater credit risk (see chart 5). The higher interest rates of loans on payment holidays is a reflection of higher risk, in our view.

Chart 5

The DTI distribution shows a similar negative pattern among loans on payment holidays: 5.1% have a DTI above 40% compared with 3.9% for loans that are not.

Each of these five variables has a positive coefficient in our logistic regression. In other words, U.K. RMBS pools comprising owner-occupied loans with a high current LTV, a high DTI, a high current interest rate, a high number of reperforming arrangements, and previous arrears (e.g. nonconforming pools), or any combination of these variables, are likely to have a higher number on payment holidays, in our view.

A few caveats apply to our analysis. First, the proportion of payment holidays is a function of the U.K.'s regulatory guidance and the initial underwriting performed by lenders. Although the regulatory guidance applies nationwide, it varies across jurisdictions, and lenders may have very different underwriting policies. Both result in a wide range of payment holiday utilization rates. Eventually, some lenders may approve payment holidays on a case-by-case basis, while others may have a more automated process.

Buy-To-Let Loans On Payment Holidays Show Similar Trends

For buy-to-let loans, the likelihood of being on payment holiday is similar to our results for owner-occupied loans. However, there are some notable trends for buy-to-let loans.

First, the debt-service coverage ratio (DSCR) and the indexed valuation are both statistically significant with a negative and positive coefficient, respectively. In other words, the lower the debt ratio and the higher the property value based on the indexed valuation, the more likely a loan will be on payment holiday, and vice-versa, all other things being equal (see chart 6).

Chart 6

Second, of these two variables, region plays a bigger role in the indexed valuations, as property values in London are more than 2.5 times those elsewhere. The DSCR for London properties is roughly the same for all others. One possible explanation may be that a more expensive property (based on its value) may require a higher loan value to finance it, which in turn will result in a greater monthly repayment and a lower DSCR.

In relative terms, the difference in indexed valuation is more substantial outside London, which may be an early sign of overvaluation of these properties and an upcoming price correction in case of future repossession, in our view.

County-court judgments also matter, and borrowers with at least one are more likely to be on payment holidays.

Our analysis of first-time buyers does not show conclusive results given the small size of our sample of first-time buyers and their concentration in a few pools.

Payment Holiday Take-Up Is Fairly Uniform Across Regions

The percentage of borrowers on payment holidays varies between 15.4% in London and 19.2% in the North West. The number of workers participating in the CJRS and SEISS ranges between 22.5% in Wales to 26.0% in West Midlands (see chart 7).

Based on these data, there are between one-third to two-thirds as many people in government schemes than borrowers on mortgage payment holidays as a proportion, which leads us to believe that those with mortgage payment holidays are probably also receiving government support. However, our universe of rated RMBS only accounts for a small proportion of the overall market, and both government-sponsored schemes target any working person, not just those with mortgages. Some workers may be debt-free or renters, which means they may be less likely to apply for either scheme or not part of our sample. That said, we would expect a high proportion of borrowers on payment holidays to be participating in the CJRS or SEISS.

Chart 7

There is no material difference between the income earned by owner-occupied borrowers on payment holidays and those not at the national level. However, there is a difference between regions because of the income and property price disparity (see chart 8).

Chart 8

Risk Layering May Lead To Pockets Of Risk

Despite the conservative guidelines outlined in the Mortgage Market Review in 2014, risk layering can still happen, in our view. For instance, some non-legacy borrowers may be exposed to loans with high current LTVs because they did not benefit from the strong property price growth that occurred in the wake of the global financial crisis until 2017. Meanwhile, legacy borrowers may have recorded county-court judgments or reperforming arrangements with a high loan-to-income ratio.

Chart 9 looks at the stratification of the population of borrowers on payment holidays across several risk dimensions. The circumference represents the current LTV (within a 2% bracket), while the radius is the proportion of borrowers (featuring specific credit characteristics within this current LTV bracket) of the overall distribution of borrowers on payment holidays.

At a high current LTV, very few borrowers are on payment holidays and in arrears at the same time as we believe lenders did not grant them any payment holidays. That said, there is a concentration of borrowers on payment holidays with a previous or current arrears balance or with a county-court judgement in the mid-70% current LTV bracket. For instance, a little over 4% of borrowers on payment holidays have a loan with a current LTV between 74% and 76% and have a county-court judgement (see the bottom left quadrant).

On the positive side, there is a slew of arrears and performance arrangements for loans on payment holidays with a very low current LTV. These loans are to legacy borrowers who have a weak credit profile but now have very little leverage. The loss severity will thus be limited in case of repossession. The most vulnerable borrowers on payment holidays, or those most exposed to risk layering, are likely to be from the most recent vintages.

Chart 9

Recent Vintages Of Borrowers On Payment Holidays Tend To Exhibit Higher Risk

For these borrowers, loans with payment holidays exhibit a fatter tail on the lower end of the seasoning distribution: 30.9% have seasoning of less than five years compared with 26.7% for those not on payment holidays. They are thus also more likely to be paying a fixed-floating rate (47.3% compared with 40.1%) and have just reverted or are about to do so for those unable to refinance.

These borrowers also display credit metrics of lower quality, such as a higher current LTV or DTI for owner-occupied borrowers (see chart 10). As mentioned, over the past five years, the intense banking competition for borrowers has narrowed the interest rate gap between low and high LTV loans while the number of mortgage products also rose for this kind of borrower. For instance, the DTI average has deteriorated over the past few years, increasing to 23.4% in 2020 from 15% in 2012.

Chart 10

Ultimate Impact Of Payment Holidays Remains To Be Seen

The overall impact of payment holidays will only become clear once lenders implement their exit strategies. Based on the FCA's guidance, there are primarily three scenarios:

• The borrower returns to full employment and can afford the mortgage payment. The borrower will only have to agree with the lender a repayment plan factoring in the suspended payments due during the payment holiday period via a higher repayment or a longer loan maturity. Anecdotal evidence shows that this is the most common and preferred scenario for lenders and presents little adverse credit impact on transactions.
• The borrower can only pay a portion of the previous mortgage repayment. In this case, the borrower can ask for a reduced payment for an additional three months. This scenario represents a maturity risk extension.
• The borrower still cannot afford any repayment. The borrower can ask for another three-month extension. This becomes a credit risk rather than extension or liquidity risk.

We believe it is still early to predict, even though some lenders granted three-month extensions at the end of June, and it may take a few more months before roll rates emerge. A consultation between lenders and the FCA is currently taking place to discuss the next steps. The combination of the third scenario, the presence of some pockets of risk, and a gloomy economic outlook, including rising unemployment and underemployment, may complicate lenders' ability to carry out exit strategies.

Even so, we think that nonconforming borrowers will be the most affected. Based on our latest macroeconomic forecasts and the historical relationship between unemployment and arrears, we could witness a sharp rise in delinquencies by mid-single-digits, especially in nonconforming transactions (see "The U.K. Faces A Steep Climb To Recovery," published July 1, 2020).

Some transactions may also experience losses because repossessions are suspended until Oct. 31, 2020, meaning that the property market could see a flurry of properties flooding the market, further fuelling a downward price correction, even though banks may be reluctant to foreclose too soon.

Finally, we expect that roll rates will be transaction-specific. Forecasting roll rates may be a daunting task because they will depend on borrowers' evolving circumstances after their payment holidays end and how lenders accommodate for these circumstances, which will also be influenced by the regulatory framework.

S&P Global Ratings acknowledges a high degree of uncertainty about the evolution of the coronavirus pandemic. The consensus among health experts is that the pandemic may now be at, or near, its peak in some regions, but will remain a threat until a vaccine or effective treatment is widely available, which may not occur until the second half of 2021. We are using this assumption in assessing the economic and credit implications associated with the pandemic (see our research here: www.spglobal.com/ratings). As the situation evolves, we will update our assumptions and estimates accordingly.

Appendix

Appendix table

Calculation methodology

We performed a backward stepwise logistic regression on a number of variables, which could potentially explain whether a loan would be on payment holiday.

All the variables described in this article are statistically significant at 99%. The predictive model was also statistically significant at the same level of confidence.

Incomes were indexed using the Office for National Statistics' wage growth data, and the net salaries were calculated using the HMRC national insurance (Category A) and tax rates for 2020/2021.

Property values were indexed using Land Registry data on a regional basis up to first-quarter 2020.

The debt-service coverage ratio of buy-to-let properties is calculated based on a hypothetical interest rate of 5.5%.

Related Research

• The U.K. Faces A Steep Climb To Recovery, July 1, 2020
• Credit FAQ: How European ABS And RMBS Servicers Are Managing COVID-19 Disruption And Payment Holidays, June 4, 2020
• Credit FAQ: Reporting Requirements For COVID-19 Payment Holidays In European Structured Finance, May 27, 2020
• More Than One-Third Of U.K. Legacy Borrowers Are "Mortgage Prisoners", Sept. 4, 2019
• U.K. Nonconforming RMBS: One In Four Borrowers Are Just About Managing To Make Their Monthly Payments, June 3, 2019
• Global Methodology And Assumptions: Assessing Pools Of Residential Loans, Jan. 25, 2019