11 Sep, 2026
Four mortgage cases this month, three approved and one stopped by the stress test
Cases
A tidy home office desk by a window with a laptop, notebook, mug of tea and a bundle of envelopes

The four cases below are the kind that reach me most weeks. They are worked examples built on this month’s published rates rather than on anyone’s actual paperwork, because the arithmetic is the part worth showing. Three get an approval. One does not.

In all four, the deciding number was not the loan to income multiple. It was the stressed monthly payment plus whatever the borrower was already paying out. The fourth failed at £2,374 a month against a £72,000 salary, on figures I should have asked for on the first call.

What do the September 2026 numbers actually look like for a borrower

The Bank of England held Bank Rate at 3.75% on 30 July 2026, on a 6 to 3 vote, with three members preferring a rise to 4%. The next decision lands on 17 September.

Moneyfacts averages at 1 September 2026 put the average two-year fix at 5.52% and the average five-year fix at 5.64%. Those are whole-of-market averages across every loan to value band, so a big deposit beats them and a 5% deposit does not. On 7 September Moneyfacts reported major lenders repricing upwards as swap rates moved again, and put the average standard variable rate at 7.13%. I have used those three rates throughout, and I have flagged where a real case would price away from them.

HM Land Registry put the average first-time buyer property at £229,107 in June 2026, and UK Finance found the typical first-time buyer spending 22.6% of gross income on mortgage payments that month, the highest share since the financial crisis.

How did a first-time buyer couple on £58,000 get to an approval

Take a couple buying at £230,000, roughly the first-time buyer average, with a 10% deposit of £23,000 and a loan of £207,000 over 30 years. Joint income £58,000.

The multiple is 3.57, comfortably under 4.5, so it never goes near the cap. At 5.52% the payment is about £1,178 a month, £14,136 a year, which is 24.4% of gross income. Above the 22.6% typical figure, and I said so. The stress test is where it tightens: at 7.13% the same loan costs £217 a month more. No car finance, no loans, one credit card cleared monthly. It went through.

Case one, £207,000 over 30 years Monthly payment Share of £58,000 gross
At 5.52%, the average two-year fix £1,178 24.4%
At 7.13%, the average SVR used as a stress rate £1,395 28.9%

Source: payments calculated on Moneyfacts average rates at 1 September 2026.

Why did a remortgage of £185,000 still cost £147 a month more

This is the case I see most often. UK Finance counted 529,750 refinancing transactions in the second quarter of 2026, up 28% on a year earlier, with about 1.8 million mortgages reaching the end of their deal rate during 2026.

Take £185,000 outstanding with 20 years left, rolling off a fix at 4.19%. The payment now is about £1,140 a month. On a new five-year fix at 5.64% it is about £1,287. Landing on the 7.13% SVR instead would cost about £1,449.

So the fix costs £147 a month more than the deal they are leaving, £1,764 a year, and saves £162 a month against the reversion rate, £1,944 a year. Both are true, and they feel very different depending on which you lead with.

One rule change matters here. The FCA’s PS25/11, published 22 July 2025, extended the modified affordability assessment to remortgaging with a new lender where the new deal is more affordable. It is permissive, so lenders choose whether to use it, which is why two can read the same borrower differently. I track who does what in my notes on rates and lender criteria.

How does self-employed income change what a lender will lend

Two years of accounts, net profit of £41,000 in the earlier year and £52,000 in the later one. Nothing unusual, just a business growing.

A lender averaging the two years works from £46,500, so 4.5 times income is £209,250. One using the latest year alone works from £52,000, so 4.5 times is £234,000. That is a £24,750 difference in borrowing capacity from one set of accounts.

On a £280,000 purchase, the averaging lender needs a £70,750 deposit and the latest-year lender needs £46,000. Same buyer, same accounts, £24,750 between a yes and a no. This one was approved, and the work was placing it correctly first time.

A row of British red brick terraced houses on a quiet street under an overcast sky

What stopped the fourth case, and what did I get wrong

A buyer at £315,000 with a 15% deposit of £47,250, so a loan of £267,750 over 30 years, on a single salary of £72,000. Loan to income of 3.72. On paper, fine.

At 5.52% the payment is about £1,524 a month, and stressed at 7.13% about £1,805. Arguably workable on £72,000 gross.

Then the commitments. Car finance at £389 a month with 31 months to run, and a £6,000 credit card balance carrying a minimum payment of roughly £180. That is £569 a month before a single mortgage payment, and lenders count it. Stressed mortgage plus commitments comes to £2,374 a month, £28,488 a year, just under 40% of gross income. It did not pass.

My part in it: I ran income and deposit on the first call and only asked for credit commitments on the second, by which point an offer had been accepted and a survey paid for. The car finance was the whole case and I found it a week late. I now take the balance and payment on every agreement before I say anything encouraging about a number.

The lender was not being awkward. MCOB 11.6.18R requires lenders to take account of likely future interest rate rises for at least five years unless the rate is fixed for five years or more. The rule did what it is there to do.

Is the loan to income cap really the thing holding buyers back

The flow limit means a lender must keep new loans at 4.5 times income or above to no more than 15% of its new lending each year, confirmed in PS11/25 on 8 July 2025. The regulators then consulted on 1 April 2026 on removing that limit for individual firms while keeping aggregate flow consistent with 15%. That closed on 1 July 2026, with implementation expected later this year.

Here is the counterweight, and it is the case against my own headline. A looser cap would not have saved any of my four. Three sat nowhere near 4.5 times, and the fourth failed on commitments rather than multiple. What it does instead is let people borrow more at a moment when the typical first-time buyer already pays 22.6% of gross income and three MPC members were voting for a rise rather than a cut.

It does bind one group hard, the high earners early in their careers in expensive areas where 4.5 times will not reach the price of a flat. It is just not the rule that stops most people.

Frequently asked questions

Does a mortgage stress test use the Bank Rate? No. It uses a rate the lender sets, and many add a margin to their own reversion rate, which is why the 7.13% average SVR works as a proxy here. The FCA says firms should reference independent forecasts of market expectations, not their own.

Is a stress test always required? No. Under MCOB 11.6.18R lenders must consider likely rate rises for at least five years, but no test is required where the rate is fixed for five years or more.

Does car finance really stop a mortgage? It can, and in the fourth case it did. A £389 agreement plus a £180 card minimum is £569 of committed monthly outgoings, taken off affordability before the stressed mortgage payment is added.

When do the loan to income changes take effect? Not yet. The 15% limit for individual firms still sits in PRA rules, the consultation closed on 1 July 2026, and implementation is expected in the second half of the year.

What I would do, on these figures, is settle the commitments question before the pricing question, every time. Total every agreement, then work backwards from the stressed payment to the loan. Case one had £1,395 of stressed room and used it. Case four had £1,805 on top of £569 already spoken for, and no rate on the market fixed that. Running versions live in the case notes, and how I work is on the about page.

Sources

The four cases above are illustrative composites built from published figures, not accounts of individual clients. This is not financial advice.