The U.K. mortgage market is characterized by a high level of "switching," according to the Mortgage Market Study published by the Financial Conduct Authority (FCA) in March 2019. In other words, residential borrowers generally borrow at a favorable discount rate for two, three, or five years before seeing their mortgage interest rate rising to a higher contractual reversion rate. Before the reversionary date, borrowers typically switch to another discount rate to avoid the payment shock. This strategy, however, is not available to everyone.

The years of poor lending practices preceding the crisis followed by the regulatory response, as well as lenders' changing credit appetite, have led to the creation of "mortgage prisoners". The FCA has identified that they, despite being up to date with payments, are unable to refinance under current affordability rules. They are trapped in their current deal.

In April 2014, the FCA amended the Mortgages Conduct of Business (MCOB) rules laying out the "responsible lending rules". The stricter lending rules aimed at preventing any lending excesses had the unexpected effect of preventing mortgage prisoners from refinancing. Indeed, the FCA estimated that 30,000 borrowers would benefit from switching but are unable to do so and that 96% of these borrowers took their mortgages or their mortgage reverted before the crisis.

In March 2019, the FCA published a consultation paper that outlines proposals aimed at helping mortgage prisoners. In particular, lenders would be permitted to apply a "Modified Affordability Assessment" (MAA) for borrowers who meet certain criteria. Moreover, lenders could disapply the "responsible lending rules" as defined in the MCOB rules (2014): income must be independently assessed, affordability under interest rate stress tests as well as the effect of future changes to income and expenditures should be analyzed, and interest-only (IO) borrowers must submit a credible principle repayment strategy.

The three-month consultation stage closed in June 2019. The FCA intends to publish the final rule changes in a policy statement in fourth quarter 2019.

This development raises a few questions:

• What proportion of the legacy borrowers in S&P Global Ratings' residential mortgage-backed securities (RMBS) pools are indeed mortgage prisoners?
• How many of them could refinance?
• Are they concentrated in some regions?
• How much would they save were they to refinance?
• What would be the impact of negative selection on the residual pools?

Our analysis of more than 400,000 legacy mortgages (originated prior to 2009) across 102 U.K. transactions addresses these questions (see Appendix 1 for the full list of transactions).

37% Of Legacy Borrowers Are "Mortgage Prisoners"

Based on our dataset, most mortgage prisoners are currently paying a high floating rate (at an average of 3.17%) or a high standard reversion rate (at an average of 4.94%) because their deal has lapsed beyond the incentivized deal period for the latter. These consumers thus have a clear incentive to refinance in order to save money, but cannot.

Defining mortgage prisoners (eligibility for the proposed MAA)

The proposed rules put forward by the FCA specify circumstances under which lenders may deviate from the existing ones. In particular, lenders will be permitted to apply a MAA and disapply the rules stated above if a borrower:

• Is the first-charge owner-occupier;
• Is up to date with payments (both presently and through the last 12 months);
• Seeks a like-for-like mortgage (i.e., does not wish to borrow more, except to pay for fees);
• Has a minimum remaining loan term of two years; and
• Has a minimum outstanding balance of £10,000.

Based on our calculations, 37% of legacy borrowers meet the eligibility criteria and are thus mortgage prisoners. We only consider legacy borrowers in our analysis as they are the main focus of the assessment conducted by the FCA in the Mortgage Market Study.

Assessing affordability

If a borrower satisfies these proposed eligibility criteria, their mortgage's affordability is assessed differently: a mortgage is deemed affordable if both the new interest rate and the monthly payment are lower than those prevailing for borrower's current deal. Our analysis takes into account upfront fees (see Appendix 2 for details about our calculation methodologies).

Chart 1 shows the prevalence of mortgage prisoners, and their risk classification, for the 10 largest transactions (by current balance), excluding transactions with a significant (greater than 25%) proportion of buy-to-let (BTL) loans.

Chart 1

We can differentiate between two types of mortgage prisoners: those who we consider "high risk" (and thus less likely to be offered a new deal by lenders for commercial reasons) and those who are not.

In our analysis, we considered borrowers as high risk by defining loan-to-value (LTV), loan-to-income (LTI), and debt service ratio (DSR) thresholds of 95%, 4.5x, and 40%, respectively.

The deal with the largest proportion (93.90%, by balance) of mortgage prisoners is Farringdon Mortgages No. 2 PLC of which less than 1% are considered risky.

The concentration (by balance) of high-risk borrowers among all legacy borrowers in the deals varies. Considering deals with less than 25% BTL loans, Stratton Mortgage Funding PLC has the lowest level of high-risk borrowers, just 0.59%, while Towd Point Mortgage Funding 2016-Vantage1 PLC has a much greater concentration (22.29%). The average concentration (by balance) of high-risk borrowers in a pool is 11.10%.

The variation in MAA eligibility between regions is relatively modest, fluctuating between 33.49% (by balance) in London and 50.06% (by balance) in Northern Ireland. This is primarily because London has a higher proportion of BTL loans than that in Northern Ireland (55.76% compared with 25.64%). Although Northern Ireland has the greatest concentration (by balance) of mortgage prisoners, approximately one-third (31.79%) of them are high risk primarily due to their current LTV (CLTV) ratio exceeding the 95% LTV threshold (see chart 2).

Chart 2

62% of mortgage prisoners could refinance

The impact of the proposed MAA on a pool would be determined not only by the size of its mortgage prisoner population, but also by the "conversion rate": i.e., the proportion (by balance) who could potentially refinance.

To refinance, a mortgage prisoner would have to pass the affordability assessment mentioned above and be perceived as low risk. On the other hand, a high-risk mortgage prisoner or one already paying a low interest rate would have a low likelihood or little incentive to refinance.

Chart 3

The average conversion rate in our universe, considering only those deals with at least one mortgage prisoner, is 61.89% but exhibits a wide distribution.

31.80% of legacy borrowers in Trinidad Mortgage Securities 2018-1 PLC are mortgage prisoners, of which 91.27% could potentially refinance.

Excluding BTL deals, the lowest conversion rate is 33.09%, that of Durham Mortgages A PLC. This is because these mortgage prisoners are paying relatively low interest rates--2.45% on average--so there is little incentive or opportunity for them to get a better deal.

There are large differences in conversion rates between regions, as illustrated in chart 4. The lowest conversion rates are in Northern Ireland (13.29%) and the North (18.27%). These regions have higher levels of high-risk borrowers among their mortgage prisoners, and those who are not considered high-risk still tend to have a high CLTV ratio.

The region with the highest conversion rate is London (73.51%). Mortgage prisoners in London are mostly low-risk borrowers and enjoy a low CLTV ratio--they are more desirable borrowers. The strong house price recovery in the South East from 2009 until 2016 help bring down the CLTV ratio.

Chart 4

But this represents only a third of legacy borrowers

Substituting the number of mortgage prisoners by the total number of legacy borrowers, we can easily derive the conversion rate (legacy conversion rate) for all legacy borrowers in each pool of our universe (see chart 3).

On average, 29.46% (by balance) of all legacy borrowers in a given transaction could potentially refinance.

Regional disparities in the number of potentially refinancing legacy borrowers are also apparent. In London and the South East, approximately a quarter of legacy borrowers could potentially refinance. In the North and Northern Ireland, where the legacy borrowers tend to be riskier and are more likely to be in arrears, a little over 5% could refinance.

Mortgage Prisoners' Credit Metrics Compare Favorably With Remortgagers

On average, mortgage prisoners in a given pool have a median CLTV ratio, LTI ratio, and DSR of 47.23%, 2.19x and 15.37%, respectively based on our dataset.

Comparing these metrics with the remortgagers in the wider U.K. mortgage market sheds some light on their relative creditworthiness: As of June 2019, U.K. remortgagers had a LTV ratio, LTI ratio, and DSR of 57.60%, 2.74x, and 16.30%, respectively, according to U.K. Finance (a trade body representing U.K. mortgage lenders). This is the evidence that mortgage prisoners would be in a favorable position to apply for a remortgage were the proposals made by the FCA implemented, as their credit profile is no worse than current remortgagers in the wider market, in our view.

Chart 5 illustrates the distribution of transactions' CLTV ratios (y-axis), DSR (x-axis), and LTI ratios (bubble size). The further to the top right and the bigger the bubble, the more leveraged the mortgage prisoners in a given transaction.

Chart 6 is similar to chart 5 but shows the distribution by region rather than by transaction. Unsurprisingly, the regions in the South are in the bottom left quadrant, while the Northern regions feature in the upper half of the chart mainly due to historical regional affordability issue and the more dynamic house price growth in the South.

The median DSR of the mortgage prisoners in any given pool in our universe who can potentially refinance is currently 12.56%. After refinancing and benefitting from lower monthly payments, this figure would fall to 11.58%.

Chart 5

Chart 6

Freed Mortgage Prisoners Would Save £108 A Month

We calculated that the average monthly payment reductions among those borrowers who could potentially refinance are 26.06% in relative terms or £108 per month in absolute terms.

Chart 7 shows the magnitude of the savings in absolute and relative terms. It calculates the average variation (expressed in percentages; right scale) between the current monthly loan repayment and the new one as well as the changes in monetary terms (left scale) should they refinance in a given transaction.

Chart 7

There is a strong positive correlation between the interest rate level, the monthly payment size, and the reduction in payment in both absolute and relative terms. This relationship is unsurprisingly stronger for IO loans than for repayment loans.

Trinidad Mortgage Securities 2018-1 PLC has the largest relative reduction in payments. This is because the average interest rate among its mortgage prisoners potentially refinancing is 5.67% despite a low median CLTV ratio of 37.52%.

The average reduction in payments varies by transaction: potentially refinancing borrowers in Durham Mortgages A PLC save £19 per month on average, while Trinidad Mortgage Securities 2018-1 PLC has greater reductions, averaging £478.15 (the average monthly payment in this transaction is £1,200). A key factor behind this is the deal's IO loans composition (69.98% by balance).

In terms of monthly payments, the average savings in our universe of transactions would be highest in East Anglia (26.45%), and the South East (26.04%). Savings are slightly more modest in the North (17.31%) and North West (21.89%) (see chart 8).

Chart 8

Repayments Would Rise, But Negative Selection Remains Limited

Increase in prepayments

The main impact on the pools of this refinancing activity would be an increase in prepayments and the resulting potential negative selection, in our view.

Increases in prepayments may change existing weighted-average life assumptions, which in turn, for some transactions currently paying pro rata, may mean a permanent switch to sequential payment sooner than currently expected.

In our analysis, we simulate a range of scenarios in which we vary the proportion of mortgage prisoners who could refinance and who actually do. This reflects the fact that consumers may not choose to refinance, perhaps due to a lack of engagement or lack of awareness of the rule changes.

Specifically, we consider three scenarios with different proportions of refinancing: one each in which all, half, or a quarter of those who could potentially refinance choose to do so. In the comments that follow, we assume the second scenario.

The average decrease in the pool factor (the current mortgage balance outstanding over the closing mortgage balance) would be 9.01 percentage points.

Chart 9

Limited negative selection

A limited negative selection could appear among the remaining pools, in our view.

The effect on the concentration of IO and self-certified borrowers is limited and varies between deals. This is reasonable, as the proposed MAA eligibility criteria would not take a view on IO or self-certification and so these borrowers would not be automatically excluded from refinancing.

However, the proposed MAA eligibility criteria would exclude borrowers in arrears. Therefore, the average arrears level in our universe would increase to 8.47% for the legacy borrowers that do not refinance from 7.48% presently.

The remaining borrowers, including some who were ineligible for the MAA, are high-risk and in arrears. There are 4,076 such borrowers in our sample. They have the poorest credit profile of all borrowers, with average CLTV ratios, LTI ratios, and DSRs of 63.38%, 4.58x, and 26.03%, respectively. They would form a "pocket of risk."

Appendix 1: List Of Transactions

Table 1

Appendix 2: Methodology

We calculated the values for the CLTV ratios, LTI ratios, and DSRs based on the current balance data and indexed income and valuation data as of June 30, 2019.

Borrowers' primary and secondary incomes were indexed using regional data from the Office for National Statistics. Only verified incomes were indexed.

The property valuation was indexed using regional data from the Land Registry.

We calculated the borrowers' after-tax incomes in order to derive the DSR. The tax rates and thresholds applied assumed the individual resides in England with a personal allowance of £12,500 in order to obtain the income tax. National Insurance contributions were calculated on the basis of Category A rates and thresholds.

The potential interest rates offered to borrowers were based on Bank of England data for the five-year fixed products.

We simulated the potential refinance deals based on a rate of 1.97% if the CLTV ratios are below 75% and 3.41% otherwise. Then we computed the "adjusted" monthly payments, after fees of £1,500 are factored in and spread across the deal period, and derived the resulting potential rate and monthly payments.

The FCA proposals stipulate that, for a borrower who is MAA eligible, a new deal is deemed affordable only if it has lower monthly payments and a lower interest rate. To determine whether this is the case, we compute the potential monthly payments and rate after accounting for the aforementioned refinancing fees.