RMBS Data Shed Light on UK Mortgage Holidays
The dust is beginning to settle in the UK economy, and we have been busy using granular RMBS data to assess the impact of mortgage payment relief (payment “holidays”), which the Financial Conduct Authority guided lenders to offer borrowers affected by COVID-19.
UK Finance reported on June 19 that at a national level, 1.9 million borrowers had been granted payment holidays – around 1 in 6 – and we expect only a modest increase from here.
Payment holidays themselves are nothing new in the UK or Europe; this is one forbearance tool at a servicer’s disposal when a borrower communicates a short term issue with their ability to pay. Our experience is that it is not commonly used, because borrowers struggling to meet financial commitments would ordinarily reduce household expenses and exhaust savings before making a call for mortgage assistance.
If we pause and look across Europe, the UK stands out on its own. The FCA’s guidance was rapid and asked lenders to ‘not have regard’ for their own commercial interests when granting payment holidays to impacted borrowers. This provided a route to a material amount of ‘cost’ support for consumer budgets. In addition, there has been a significant amount of support on the ‘income’ front through the Job Retention Scheme, and similar support for the self-employed. There was a slight delay in income support landing in bank accounts, which may have prompted the push for payment holidays, but undoubtedly the UK has done plenty to aid consumers on both income and costs.
Over in Europe, existing labour market support mechanisms in many countries including Germany, France, Italy and the Netherlands were utilised to provide income support typically for over 80% of salary for almost a year (there are nuances between jurisdictions but we consider the UK ‘furlough’ scheme to have broadly mirrored this). However, the European income support has resulted in a far smaller emphasis on offering payment holidays and leaves us contemplating why the FCA published such strong guidance to lenders. The Netherlands, for example, has just 19,000 payment holidays.
As an aside, there is a question over the potential moral hazard that the FCA has introduced in its haste to roll out the scheme. On the one hand, lenders have been guided to provide financial relief and not ask questions, moving away from traditional processes for assisting stressed borrowers, and we suspect they will also fear the reputational risk of foreclosing borrowers at the end of payment holidays (Lloyds Bank was not alone in running a national TV campaign almost inviting borrowers to request one). On the other hand, borrowers can take a payment holiday without the provision of data as to their circumstances and without risk of impairing their credit standing. Undoubtedly this continues to be a very tough period for many families, but by offering a “free option” with no checks, the government may have introduced some unintended consequences.
For example, there is a likelihood that some borrowers could take a payment holiday for six months and then find themselves made redundant, unable to pay further and therefore already effectively six months in arrears. Lenders will undoubtedly feel obliged to offer as much further help as they can, particularly as up to that point they have been instructed by the regulator not to impair the borrowers’ credit record. Inevitably though some of these borrowers will not be able to meet any further payments and ultimately the lender will have to undertake foreclosure measures, but on a loan which is subsequently in a far worse position than it would have been in the normal course.
Back to RMBS data, one consequence of the speed of response was to leave lenders to interpret for themselves how to implement the move, and we think this has contributed to the wide dispersion of payment holidays across UK RMBS deals.
Having crunched the numbers, we see four key trends:
- Self-employed borrowers, whose income is more prone to volatility, have been more severely impacted.
- Seasoned loans, particularly those originated before the GFC, perform better than recently originated loans – we attribute this in part to lower average loan sizes and thus smaller monthly payments
- Buy-to-let loans generally perform better than residential, in part because of their dual recourse nature (tenant rent and landlord)
- Prime bank or building society lenders with broadly similar underwriting standards mirror the national average of payment holidays for 1 in 6 borrowers
However, it is among the challenger banks and non-bank lenders who commonly compete for non-high street borrowers where large discrepancies arise, and here we see a bigger range from the mid-teens to 40%. These deals typically comprise a higher number of younger and self-employed borrowers, first-time buyers, and those with weaker credit histories.
While the numbers here may sound alarming, it is important to state that we do not consider this poses a credit risk for RMBS bonds using our own revised scenarios. Deals with a perceived weaker average borrower naturally start life with more interest coverage and loss protection, but we also expect a good number of affected borrowers took payment holidays as a free option to provide themselves with a buffer, and will therefore recommence payments in an orderly manner. Finally, some high street lenders such as Santander are providing support to deals and absorbing the missed revenue.
The loan-level data available to ABS holders can be a very useful tool for judging the impact of an economic shock on the real economy. As it stands lenders have taken the FCA’s somewhat forthright guidance in their stride and we do not expect the relatively high level of payment holidays in the UK to present a credit risk for RMBS.
July will be key as this is when the first batch of ‘holidayed’ loans are scheduled to resume payments, albeit some borrowers may choose to extend their holidays, as the FCA have allowed for, and we will also see if lenders will give borrowers a clean slate by capitalising missed amounts as expected. Watch this space.