Purchase approvals fell. Remortgaging did not. That is your July.

Net borrowing roughly halving in a single month is the headline, and it is the least useful number here. Net lending is noisy and one month is not a trend.

The approvals split is the useful part, because approvals are forward-looking — they are work that has not yet completed — and because the two lines moved in opposite directions.

Why the divergence matters more than either number

Purchase business depends on somebody deciding to move. Remortgage business depends on a date arriving. When rates rise and confidence softens, the first is postponed and the second cannot be.

A broker whose pipeline is weighted to purchase feels a month like July as a slump. A broker with a book of maturities feels it as an ordinary month with more phone calls. Same market, two different Julys, and the difference is not skill.

The effective rate moving from 4.35% to 4.45% on newly drawn mortgages is the mechanism sitting under both lines. Ten basis points is not dramatic in itself, but it is the direction that changes the conversation with a client whose fixed rate ends in the spring.

What this implies operationally

If remortgaging holds while purchase softens, the constraint on a brokerage stops being lead generation and starts being throughput. The work is already in the book; it arrives on known dates; the question is whether you reach each client early enough and whether the case takes five hours of chasing or two.

That is exactly the arithmetic in where the hours go on a case — fifteen to twenty rounds of messages, upwards of five hours on the broker-to-lender leg, and a thirty-five per cent chance of corrections. In a purchase-led market that cost is hidden by the deal flow. In a maturity-led one it is the whole margin.

It also puts weight on knowing your own maturity calendar precisely, which is a data question rather than a sales one, and the sort of thing a pipeline answers better than a spreadsheet somebody updates on Fridays.

What this does not tell you

One month is one month. July is a holiday month, approvals are volatile, and the Bank revises. Do not build a strategy on a single print, and be suspicious of anybody who does — particularly if they are selling you something that solves it.

It also says nothing about your local market or your client base. National approvals can fall while a specific segment rises, and a broker working with self-employed or complex income is not sampling the same population as the aggregate. The FCA's own work on exactly those borrowers is running now, which is the subject of the mortgage rulebook in parts.

The one thing we would take from it without hesitation is the divergence, because it is structural rather than seasonal, and because it points at the part of a brokerage that responds to effort. Our work with regulated firms starts at the case file for that reason.