The buy-to-let sector and financial stability

Falling house prices following the 1990s recession led to a significant number of households losing their homes. This led to a surge in rental demand, which the socially rented sector was not able to meet. The Housing Act of 1988 introduced legislative changes, in particular fixed-term occupancy, which helped to increase PRS investment and meet the demand for rental properties.

BTL mortgage products were initially introduced in 1996, and the sector expanded substantially in the early 2000s. By around 2007, it was about 25% to 40% cheaper to rent a home than to service a mortgage, as many households were priced out of the housing market by a combination of record high house prices and subdued wage growth. Between 2000 and 2008, the outstanding BTL mortgage stock increased from £9 billion to £140 billion – roughly 12% of the overall mortgage market (Chart 1).

Chart 1: Evolution of the UK mortgage market over time

Footnotes

• Sources: ONS, UK Finance industry tables and Bank calculations.

Since 2015, the BTL market has grown at a slower pace. Annual growth of the stock of loans has averaged around 5%, compared to 8.5% between 2008 and 2015. The current stock of BTL mortgages amounts to around £300 billion of outstanding debt, which is around a fifth of the total mortgage market.

3: Why does the BTL market matter for financial stability?

Pressures in the BTL market can affect financial stability through two main channels:

• Lender resilience channel, where lenders experience losses when highly indebted households face repayment difficulties. Lenders’ losses could stem from BTL mortgages directly, or from other credit held by landlords or renters.
• Borrower resilience channel, where highly indebted households are stretched by higher rents or mortgage payments, higher housing maintenance costs, or falling asset values. In the context of the BTL market, the borrower resilience channel can be transmitted through both landlords and renters. These pressures may lead households to cut spending sharply or lead landlords to sell.

These channels interact with each other and can amplify shocks, especially if pressures led many landlords to decide to exit the market at the same time.

Recognising the potential for financial stability risks to arise from the BTL sector, the FPC has Powers of Direction to impose limits on low interest coverage ratio (ICR) and high LTV lending in the BTL market extended by regulated entities (Table A).

Moreover, the PRA supervisory statement 13/16 – Underwriting standards for buy-to-let mortgage contracts, September 2016 (SS 13/16), requires PRA-regulated lenders, when assessing the affordability of new BTL mortgages, to apply either or both:

• Income affordability testing at a stressed interest rate of at least the higher of 5.5% or a 2 percentage point increase in BTL mortgage rates. This test is used when lenders are taking account of a borrower’s personal income as a means to meet mortgage payments, which is defined as top slicing.
• Interest rate affordability stress testing (ICR testing) to ensure that rental income is sufficient to pay interest costs, after accounting for maintenance and other costs of letting. Lenders tend to test affordability at a minimum ICR of 125%, assuming stressed interest rates as above, and the basic rate of income tax. For higher and additional-rate taxpayers, the equivalent ‘breakeven’ ICR is about 167%. In practice, lenders have tended to test those borrowers to 145%.

These testing requirements do not cover all BTL lending. They exclude lending with a fixed-rate period of five years or more and lending by non-banks, which together comprised about 75% of BTL new lending in 2023 Q3.^{footnote [1]} However, banks also apply their own affordability testing standards outside of the SS 13/16 requirements.

4: What happened in past downturns?

The BTL market experienced a sharp increase in distress during the GFC.

The three-month arrears rate for BTL mortgages increased six-fold compared with two-fold for owner-occupier mortgages (Chart 2).

Chart 2: BTL arrears rates increased significantly during the GFC ^(a)

Footnotes

• Source: UK Finance industry tables.
• (a) Chart shows the share of loans in arrears worth at least three months of repayments.

One reason for the relatively high arrears rate among BTL loans was strong lending at high-LTV ratios. Financial Services Authority analysis on post-GFC arrears data found that landlords with origination LTVs in the 90–95% segment were more than three times more likely to default than residential borrowers in the same LTV band. This may reflect that owner-occupiers are more invested in meeting their mortgage payments than landlords, given the risk of losing their home in distress. This also results in typically higher interest rates for BTL than for owner-occupier mortgage loans. More broadly, pre-GFC underwriting standards for BTL were weak in a number of ways, including weak income verification, significant lending to people with low incomes, and inadequate affordability testing.

During the Covid downturn, however, the BTL market was significantly more resilient. In particular, arrears increased only marginally over 2020 before recovering to record lows. Arrears have picked up since the start of the current monetary policy tightening cycle, but the increase has been contained so far. Much of this can be explained by improved underwriting standards, which support borrower and lender resilience. The next section covers these standards and their effects.

5: What is the current state of the BTL market?

The quality of BTL lending has held strong since 2015.

Formalising underwriting standards through SS 13/16 has helped to ensure that the quality of new BTL lending has held strong since 2015. ICRs have improved significantly between 2015 and 2022, as shown in Chart 3. The median ICR at origination stood at almost 300% in 2022, up from 242% in 2015. Following sharp increases in interest rates since end-2021, the ICR distribution on new lending has shifted significantly lower, reflecting pressures on the market. But we expect ICRs to recover somewhat going forward, as there is evidence that landlords are passing on costs to renters. And many landlords have significant ICR buffers. For example, 83% of new BTL mortgages taken out in 2022 had an ICR greater than 200%, well above ‘breakeven’ levels.

The majority of all new loans have LTV ratios below 75%. In the first three quarters of 2023, only around 1% of all new BTL loans originated with LTVs at or above 80%. These LTVs are much lower than is usual for owner-occupier mortgages. But given the majority (82%) of BTL loans are on interest only terms, rather than repaying the principal, the LTVs on BTL loans fall more slowly over time, driven foremost by rising property prices.

Chart 3: Distribution of ICRs on flow of new BTL mortgages: 2015, 2022, 2023 Q3

Footnotes

• Sources: FCA Product Sales Data and Bank calculations.

Prudent underwriting standards on new lending, alongside strong house price growth, also means that the stock of BTL mortgages has become more resilient to house price falls. The share of mortgages with LTV ratios of at least 75% fell from around 12% at the end of 2016 to around 6% in 2023 Q2 (Chart 4). The resilience of the stock means that even if prices fell by around 20%, only around 1% of mortgages would go into negative equity with the outstanding principal loan worth more than the value of the asset. If prices were to fall by around 30%, as modelled in the 2022/23 annual cyclical scenario (ACS), the number of loans falling into negative equity would increase to around 16%.

Chart 4: LTV distribution on stock of BTL mortgages

Footnotes

• Sources: Stress Test Data Framework and Bank calculations.

The composition of lenders involved in the BTL market is also important for financial stability. The seven largest UK lenders^{footnote [2]} extend the majority of BTL loans, responsible for around half of new lending. Specialist lenders and smaller banks add another 22%, building societies account for 15%, while non-bank lending accounts for 13% (Chart 5).

Chart 5: New BTL lending in 2022, by lender type

Footnotes

• Sources: UK Finance industry tables and Bank calculations.

Historically, BTL landlords have operated on a small scale, but the sector has been consolidating…

Across the PRS as a whole, just under half of landlords own only one rental property, according to the 2021 English Private Landlord Survey. At the other end of the spectrum, nearly half of private renters have a landlord who owns at least five properties (Chart 6). The great majority of landlords operate as private individuals rather than companies, especially if they own only a few properties. However, the market has been consolidating as landlords with only a few properties are less likely to enter the market and more likely to exit the market. And recent market intelligence notes that larger landlords are likely to be buying up some of the properties sold by smaller landlords. Nevertheless, there is evidence that the PRS has been moderately declining overall, although measuring changes in the size of the PRS is challenging (Box A). These trends reflect the regulatory and economic pressures facing BTL landlords.

Chart 6: Size of landlord portfolios

Footnotes

• Source: 2021 English Private Landlord Survey.

…as landlords face a number of economic and regulatory pressures.

Higher operating costs, driven by interest rate increases and the prior removal of mortgage interest tax relief (MITR), are the largest headwinds on BTL profitability (Table B). Relative to owner-occupier mortgages, the BTL market is particularly exposed to higher interest rates, because of the large proportion of interest-only mortgages in the stock of loans – 82% for BTL compared with 11% for owner-occupier loans.

Some one-off transaction costs, like reduced capital gains exemptions and increased Stamp Duty, also reduce landlords’ profits and market turnover. These costs pose relatively modest headwinds on capital gains for existing BTL landlords, given strong house price growth in recent years. However, such costs may deter new landlords from entering the market, especially given higher returns on lower-risk assets.

Renting households tend to have lower incomes on average than other households and spend a significant portion of their incomes on rent.

At £26,500, the median renter’s gross income is less than half that of the median mortgagor (£59,000). About a quarter of renters report spending more than 37% of their gross income on rent (Chart 7). Among private renters, the median income is around £35,200, and the mean housing costs as a share of income is 28%. Renters also tend to have much lower savings to draw on than other households. Nearly 28% of renters reported having no savings in the NMG 2023 H2 survey, and about 50% reported having less than three months’ worth of rent in savings.

Chart 7: Renters spend more of their income on housing compared with mortgagors ^(a)

Footnotes

• Sources: NMG Consulting survey and Bank calculations.
• (a) The chart shows the distribution of mortgagors’ and renters’ spending on housing costs, as a share of gross income. These measures respectively take rent and mortgage costs (including principal repayment), divided by total gross household income. For mortgagors, note that this measure does not capture other housing costs related to ownership, such as maintenance and building insurance.

6: How do pressures on the BTL sector matter for financial stability?

Landlords’ responses to economic and regulatory pressures have implications for financial stability.

Figure 1 maps the options that landlords have – to raise rents, absorb costs, or sell and exit the market – and their respective financial stability implications. The remainder of this section expands on the implications for lender resilience and borrower resilience, among both landlords and renters.

Figure 1: Mapping BTL pressures to financial stability risks

Lenders are resilient to the pressures in the BTL market…

Robust LTV ratios mean that lender portfolios are likely to be resilient to even severe house price falls, as noted in the 2022/23 ACS results. And, for now, arrears have remained low. In Q3 2023, the three-month BTL arrears rate was 0.6%, up from 0.5% in Q2. The equivalent owner-occupier rate was 0.9%, up from 0.8% in Q2.

Overall, financial stability risks from changes in funding conditions are limited, because the majority of lending is undertaken by traditional lenders (Chart 5). But some non-bank lenders could be vulnerable to funding risks in the residential mortgage-backed securities (RMBS) market. Around 9% of the BTL stock of loans is securitised, compared with around 5% for owner-occupiers. The RMBS market is sensitive to changes in economic and market conditions, and in investor appetite. As such, tighter financial conditions could restrict access to funding and/or drive-up funding costs for some lenders, especially those relying on securitisation.

…and spillovers to the broader housing market are contained…

Economic and regulatory pressures may lead BTL landlords to sell their properties. Given strong house price growth in recent years, many landlords would stand to make a profit by selling their properties. And a higher interest rate environment has increased the return on other safe invests, which landlords could reinvest in. A range of sources report landlords selling up, resulting in a moderate decline in the size of the PRS. However, as Box A sets out, these changes are difficult to measure.

Nevertheless, the high share of five-year fixed-rate BTL mortgages reduces the risk of a significant share of BTL properties being sold at once, because the impact of higher interest rates is staggered. The share of five-year fixed lending has increased significantly since 2014, likely reflecting mortgagors seeking to lock in low interest rates. In addition, long lead times for regulatory changes give landlords time to adapt to the changing economics of the PRS.

Even if many landlords did sell, we judge that the house price impact would be limited, partly reflecting that properties with BTL mortgages make up only around 9% of the overall UK housing stock by number. Using an agent-based model of the UK housing market we simulate a sharp 30% fall in the number of landlords, which leads to house price falls of just 4% in the long run, compared to a benchmark house price path.

…but borrower resilience is being tested.

Many landlords could absorb pressures by taking a hit to their profit margins and may be motivated to hold onto investment properties. Not all landlords will face pressures. Increased operating costs, mainly through the reduction of MITR, will only affect BTL landlords, not those who own rental properties outright. Rental properties may become more profitable for landlords who own their properties outright, particularly if they are able to raise rents in line with BTL landlords’ costs, and demand for rentals remains strong. Even among BTL landlords, there is a wide distribution of ICRs – indicating that some BTL landlords have comfortable profit margins.

But some BTL landlords are operating on thin margins, as Chart 3 illustrated. The gross rental yield on BTL investments has averaged around 6% since 2014. But net rental yields (which account for property running costs and mortgage costs) have decreased significantly since mid-2022. For an average BTL property,^{footnote [3]} the net rental yield was around 3.5% in Q2 2022. Since then, higher mortgage rates imply an average increase of around £2,500 in annual interest payments for BTL landlords. Full pass-through implies a net rental yield of 2.5%.

Rather than absorb costs, landlords could seek to raise rents further. There is a risk that landlords cut consumption, run down savings, or default on other obligations, in order to service their BTL mortgage – but we judge this to be unlikely. In the first instance, landlords are likely to try to pass some costs on to renters. UK private rents increased by 6.1% in the 12 months to October 2023, the largest annual change since the series began in 2016. Rents on new lets increased significantly more – by 10.5% in the UK in the year to September 2023, according to Zoopla. And around 13% of renting households reported experiencing rent increases above 20% in the 12 months to September, according to the NMG 2023 H2 survey. In the near term, higher rents are likely, given rising mortgage costs and strong demand.

However, there are already some signs that some renters have limited capacity to absorb further rent rises. For example, the NMG survey shows that, among surveyed private renters, the share who have been in any rental arrears at some point in the past 12 months has increased steadily since 2020 from 10% to 14% (Chart 8). And in the latest survey, 2023 H2, renters that had faced larger rent increases were more likely to report having been in arrears.

Chart 8: Share of private renters reporting having been in arrears in the past 12 months

Footnotes

• Sources: NMG Survey 2023 H2 and Bank calculations.

Adapting to higher rents may require difficult choices for renters. While many households will have benefitted from recent strong average income growth, helping them to keep pace with rising living costs, not everyone will have. Renters whose living costs are rising faster than their incomes could cut consumption, run down savings, turn to credit, or default on other financial obligations. Any of those options would reduce their resilience to future financial shocks – and, as set out above, renters already tend to be less financially resilient than other households. It may also make it harder for prospective first-time buyers to save a deposit.

Higher rents will erode the financial resilience of many households but are unlikely to pose financial stability risks. Pressures on renters could pose financial stability risks if renters responded by (i) defaulting on loans at a scale that threatened lenders, or (ii) cut back on consumption sharply, in a way that amplified an economic downturn. But neither of these risks is very likely, because:

• Banks have limited exposures to renters. By definition, most renters do not have secured loans – which make up the vast majority of banks’ retail exposures. And supervisory intelligence suggests that banks are maintaining prudent affordability standards on the flow of new lending.
• Renters have less scope to cut consumption of non-essential goods sharply. While renters’ non-housing consumption is sensitive to economic pressures, renters tend to spend less on non-essentials than other households to start with. On average, private renters spend £246 per week on non-essentials, compared to £177 for social renters and £384 for owners. By simple calculations, renters account for about 23% of non-essential expenditure (14% by private, 9% by social renters), and slightly more than 30% of overall expenditure (20% by private, 12% by social).

Box A: Is the private rented sector shrinking?

This box describes a range of measures of the size of the PRS and explains why having accurate measures is important for judging future vulnerabilities. In the rental sector, lower supply is likely to contribute to higher rental inflation. And in the housing market as a whole, large exits could put downward pressure on house prices – especially if they happen rapidly.

There are no available measures that fully capture the size of the PRS. Existing measures predominantly focus on outflows, rather than capturing the net flow. This means they may overstate outflows from the PRS, if considered in isolation, by not accounting for inflows or inter-landlord transactions.

Existing measures include:

1. Homes listed for sale within three years of being advertised for rent (Paragon).
2. Implied changes in the BtL mortgage stock (UK Finance).
3. Number of second homes subject to capital gains tax (HMRC).
4. Real estate agents’ own record of the share of properties bought and sold by investors, applied to HMRC property transactions to estimate whole market numbers (Hamptons). This is the only external estimate to include a net balance, rather than focusing solely on outflows.

Measuring changes in the size of the PRS is challenging. It often requires using different property-level data sets, which can be hard to match. And distinguishing between properties moving between the private rental and owner-occupier markets versus properties changing hands between landlords is important. This is made more difficult when properties have complex or incomplete rental histories.

We have developed a new measure of net additions to the stock of rental properties in the PRS. Our measure:

• matches HM Land Registry property transactions with rental advertisements provided by Zoopla-WhenFresh to identify properties that have been sold and subsequently rented (and vice versa). Rental advertisement data are daily UK rental listings sourced from real estate agents’ own systems, while HM Land Registry transactions are a record of all properties sold in England and Wales; and
• categorises these matches as inflows or outflows from the PRS, or inter-landlord flows, as shown in Chart A and defined as:
• owner-occupier to landlord inflows are properties advertised for rent within four years of being purchased, which have not been advertised for rent previously;
• landlord to owner-occupier outflows are properties sold within four years of being advertised for rent, which have not been advertised for rent since being sold;
• inter-landlord flows are either:
• properties advertised for rent within four years of being sold, which have been advertised for rent previously, or,
• properties sold within four years of being advertised for rent, which have been advertised for rent since.

While imperfect, our measure captures net flows across a significant share of the private rental market. The time lag between sale and rent means that observations in the most recent four-year period are particularly subject to revision (indicated in Chart A with dashed lines).

Our measure indicates modest net outflows from the PRS in aggregate, and across UK regions and property sizes.

• Landlord outflows and inflows converged over 2016 to 2020. Outflows appear to have outstripped inflows for at least the past two years, but at a more modest rate than other measures suggest. These trends are broadly consistent across regions and property size. The post-pandemic peak in outflows was more notable in London and the South East than other UK regions.

Chart A: New measure shows moderate market shrinkage ^{(a) (b) (c) (d)}

Footnotes

• Sources: HM Land Registry, Zoopla-WhenFresh and Bank calculations.
• (a) PRS flow measure covers properties sold in England and Wales. Dashed lines are subject to revision when new data is released within four years of date shown, due to the time lag between sale and rent. Underlying data start in November 2008. Data excludes demolitions, and may not capture build-to-rent. Data up to 31 December 2022.
• (b) Owner-occupier to landlord: properties sold, then listed for rent within four years that have not been listed for rent before.
• (c) Landlord to landlord: properties sold, then listed for rent within four years that have been listed for rent before. Or, properties listed for rent, then sold within four years that have been listed for rent since.
• (d) Landlord to owner-occupier: properties listed for rent then sold within four years that have not been listed for rent again.

Chart B: Comparison with UK Finance and Hamptons ^{(a) (b)}

Footnotes

• Sources: Hamptons, HMRC, HM Land Registry, UK Finance, Zoopla-WhenFresh and Bank calculations.
• (a) For Hamptons data see text on existing measures in Box A.
• (b) UK Finance number of BTL mortgages for house purchase (excluding remortgages). Properties transferred to BTL incorporations count as a house purchase.

As with any estimates of PRS flows our measure is not without limitations, including:

1. Incomplete sample – we can only analyse rental properties that are matched across the HM Land Registry and Zoopla-WhenFresh data sets.
2. Risk of overstating outflows – properties sold to a landlord but not yet listed for rent could be reclassified as inflows or inter-landlord flows once they’re listed for rent, which means outflows could be overstated. This is a potential explanation for why the combined total of inflows and inter-landlord flows diverges from the flow of new BtL mortgages from 2020 after having tracked closely before then (Chart B). Nevertheless, the trend of declining inflows is apparent in our measure, as well as the UK Finance and Hamptons measures.
3. Imperfect transaction classification – some properties have complex transaction histories or incomplete rental advert data which can lead to misclassifications. For example:

• Properties listed for rent and then sold multiple times in succession are classified as landlord to owner-occupier outflows for all sales, instead of just the first sale. This could overstate outflows.
• Properties sold multiple times between rental adverts are classified as inter-landlord transactions. This could overstate the share of inter-landlord transactions and understate inflows, if properties are not tenanted while sold.