Sunday Supplement - 9 Aug 26 - The Dead Cat Examination, Oil's Round Trip, and a £1.74bn "Profit"
Book workshop tickets: tinyurl.com/pbwoct26
"The first lesson of economics is scarcity: there is never enough of anything to fully satisfy all those who want it. The first lesson of politics is to disregard the first lesson of economics." - Thomas Sowell
The quote pertains to the Deep Dive, which this week is a declared construction special - four fresh reports, one question. The question is whether Thursday's construction PMI bounce is the start of something or just a dead cat with good hang time, and the answer runs straight through Sowell's two lessons: a trade body proposing a new Help to Buy scheme alongside some encouraging movements upwards in housebuilder shares; a think tank doing honest arithmetic on the government's £39 billion grant programme, a warranty provider counting registrations, and a results season in which the housebuilders tell us what they actually believe with their land budgets rather than what they say in the outlook statements. Scarcity, politics, and the gap between the two. We will get there.
As the Budget countdown ticks through 80 days and the market keeps rewarding the well-capitalised while quietly showing everyone else the door, our next Property Business Workshop is live and tickets are selling. Thursday the 1st of October, in Manchester, with Rod Turner and myself - and the subject this time is joint ventures and M&A. This is about how deals actually get done when solo capital is tight: structuring JVs that survive contact with reality, buying businesses and portfolios rather than just houses, and the due diligence that separates a bargain from a liability with a nice brochure. Super Early Bird pricing is running at better than 20% off, and the VIP dinner the night before is filling fast, as it always does. There are only 2 VIP dinner tickets left as I write this. Book in before you miss out: tinyurl.com/pbwoct26
Welcome back to Trumpwatch. Last week the scoreboard read six hawks across two central banks in two days; this week the data and the oil market spent five sessions arguing about whether those hawks have a case at all. We got a war that was locked and loaded on Friday, called off by Sunday, negotiable by Monday and duplicitous by Monday afternoon (TACO-standard, the analysts might say); a US jobs report that went negative for the first time in months; and an oil price that gave back most of July's terror premium inside a week. Busy, for the silly season. Let's take it in order.
Iran first. On the 1st of August the President declared the US "locked and loaded and ready" to hit Iran at levels of - his capitals, not mine - Military Terror not seen since the Second World War. By Sunday the major strikes had been called off, reportedly at the urging of the Gulf states, and by Monday we had newly planned talks billed as Iran's "last chance", with the agenda covering the reopening of the Strait of Hormuz and the nuclear file. Within about 24 hours of voicing optimism that a deal was close, the President was back on Truth Social declaring the Iranian leadership "unbelievably duplicitous" after Tehran denied talks had resumed at all. If you have lost count of the threaten-cancel-negotiate-rage cycles since April, you are in good company - I make it at least five, and I would not stake much on the count. No human and barely any AI can keep up with counting the number of “untrue Truths” - a special Trumpism - that have been knocked out over the last few months. The pattern matters more than the instalments: each cycle has ended short of the maximal strike, and the market has started pricing the pattern rather than the headlines.
Which brings us to the important development of the week, and it happened in the shipping lanes rather than the situation rooms. Iran and Oman agreed a temporary shipping route through the Strait of Hormuz - explicitly not a full reopening, Tehran was at pains to say, but a corridor. Oil traders did the rest. Brent, which spent late July north of $100 and gave this publication a headline a fortnight ago, went through three straight sessions of losses and sits in the low $80s as this goes to print - $83, a round trip of getting on for a fifth in barely a week. I will hedge the durability of that move in the hardest terms: the Royal Navy reported explosions near a tanker transiting the strait this week, the Houthis claimed an attack on a Saudi tanker, and a corridor that can be opened by agreement can be closed by one incident. But the direction of travel matters enormously for the UK story, because the entire hawkish case at the MPC - the three votes for 4% - rests on energy prices staying elevated long enough to leak into wages and expectations. A month of Brent in the $80s takes a lot of air out of that argument. We'll see whether it gets the month - it’s far too early to call. Last week I noted all we could guarantee was volatility - this week we’ve had it to the downside.
Then Friday's US jobs report, which is the sort of print that changes conversations. Nonfarm payrolls FELL by 23,000 in July against expectations of roughly 80,000 added - the first negative month in a while - and the two prior months were revised down by a combined 103,000. May now reads 63k, June just 20k. The unemployment rate "improved" to 4.1%, and here is your weekly reminder that a falling unemployment rate is not always good news: it fell because labour force participation slid to 61.4%, a five-year low. Those who remember that I often quote the 75% figure in the UK - by the way - of participation - shouldn’t compare the two - in the US they quote this from 16+, in the UK it is 16-64s, so compare the two. The denominator did the flattering, not the economy - the same arithmetic trick I bang on about with our own inactivity figures, performed transatlantically. Last week in this very publication I said Friday's report would be the first proper test of whether Warsh's three hawks have the data on their side. They do not. A committee that voted 9-3 with the dissents pointing at a hike, eight days before payrolls went negative with a hundred thousand of downward revisions attached, is a committee that will be re-reading its own minutes with some discomfort. And given what happened to the last BLS commissioner when the revisions embarrassed the White House, keep half an eye on whether the data itself becomes the story again. I'd rather it didn't; institutional statistics only get destroyed once. What you’d have expected is for yields to drop when this report came out - and they did, feeding through into a little drop in gilt yields too given the relationship between the US and UK interest rates (although this is one of those times where divergence is more likely than usual, even though the war is indeed inflationary for both parties).
The tariff machine, meanwhile, grinds on regardless. Twenty-five states are now suing the administration over the latest round of levies - the Section 301 wave that replaced the tranche the courts struck down - and the President spent part of the week demanding that tariffs on Iran be bolted onto the bipartisan Russia sanctions bill. The US imports approximately nothing from Iran, so the measure is symbolic; the risk, as trade folk pointed out within hours, is that the symbolism sinks a bill that otherwise had the votes. Tariffs as a universal adjective rather than a policy. Standard Fare.
Why does any of this matter to a portfolio in Birmingham or Bradford? Because this week it mattered in your favour, for once. Softer US data plus a deflating oil premium meant Treasury yields eased, and gilts were dragged along in the happier direction - the full damage report, or rather repair report, is in the Gilty/not Gilty section below. The imported-inflation era cuts both ways: when the world's anxiety premium deflates, so does ours. Phew - stepping away from the transatlantic theatre, back we go to the comparative sanity of the UK real-time property market.
As is customary, Chris Watkin has been relentlessly crunching the portal numbers and publishing them at Property Industry Eye. His analysis for Week 30 of 2026 - the week ending the 2nd of August - is where it is at, as always. If you want to know how the macroeconomic gridlock translates to the local high street, look no further. And a declaration of interest this week: I am actually ON the show, alongside Chris, digging through the national numbers and then a properly forensic case study of Maidstone's estate and letting agents. Always an enjoyable experience - geeking out on the stats for an hour with the UK Property Market Stat Master. Chris framed the week's question as whether we are looking at a "Burnham Bounce". My honest answer on the show, and here: I'm not convinced the new Prime Minister has much to do with it - but something did firm up in July, and it deserves a proper look rather than a slogan. Was it part of the World Cup distraction recovery? The best way to tell will be when we see retail and spending figures for July too.
Supply first. 32.8k new listings this week - fractionally below the 33.8k long-term Week 30 average, which is late July doing what late July does. The year-to-date picture: 1.102m homes listed, essentially level with 2025 (0.2% behind), 3.9% ahead of 2024, and still 11% above the 2017-19 pre-Covid norm. Regular readers will remember the ready reckoner sits at 12.5% more stock than a normal market, officially, and that I flagged a fortnight ago that the gap to normal was drifting down as summer weeks come in at seasonal levels rather than above them. That drift continues - the flow has normalised. The stock has not: 767k homes on the market on the 1st of August, against 760k a month earlier, 763k twelve months ago and 637k three years ago. The bath has stopped filling faster than it drains; it remains, as I said last week, a rim-lapper - but it is just at a steady pace right now.
Demand, and the number Chris and I spent the most time on: 24.7k homes sold subject to contract in Week 30, up from 24k the week before, against a ten-year Week 30 average of 26.1k. Year-to-date gross sales stand at 740k - 7.1% behind 2025's exceptional run, but only a whisker (0.2%) behind 2024, 10.8% ahead of 2023, and 6.5% above the pre-Covid 2017-19 norm. The net sales line is the one I'd frame: 19k net sales this week, and 575k for the year to date - which is the SAME as 2024 at this point, to the thousand. Strip out the noise and the market is transacting at decade-average volumes while carrying a fifth more stock than it is used to. That is not a boom and it is not a bust; it is a queue. The Five Ds - death, debt, divorce, downsizing and the diddy ones - keep feeding realistic vendors into the front of that queue, and the market keeps paying them properly while everyone else waits.
The friction numbers tell you who waits. 23.1k price reductions this week; 13.7% of all homes on the market were reduced in July, down a touch from 14.3% in June but still well above the 11.2% long-term average. The gap between the average asking price of everything listed (£409k) and the average asking price of what actually goes sale-agreed (£358k) sits at 14.2% - narrower than the intergalactic 16-17% long-term average, which reads to me as realism slowly winning, one reduced listing at a time. When we get into the summer holidays, history dictates that only those who are motivated bring the stock to market, and the more expensive households are also holidaying which affects averages - this happens every “normal” year. The sell-through rate: 14.2% of homes on agents' books went sale-agreed in July, up from 13.8% in June, against a pre-Covid norm of 15.5%. And the sobering pair: 46.6k exchanges recorded for July so far (this figure always climbs through August as the reporting catches up - past form says it lands in the mid-to-late 70ks) against 43.5k withdrawals, meaning only 51.7% of the homes that left agents' books last month did so by exchanging and completing. The seven-year average is 57.6%. Nearly half of leavers left unsold. Fall-throughs, at least, are behaving: a 24.3% fall-through rate against a 24.5% decade average, and just 5.07% of sold-STC homes fell through in June, below both last year's average and the ten-year norm. Average time to sell: 76 days from listing to sale agreed, and 122 days from sale agreed to completion. Six and a half months, door to door, for the average correctly-priced house. Plan accordingly.
The price-per-square-foot series - the one that front-runs the Land Registry index by five months with 98% accuracy, which is why we track it. The £345.41 for July, at 1.2% up on the year and 11.9% up on five years - much lower than some might expect on the 5 years - is a haircut from June’s £350+/psqft, but this is a relatively typical trend for the two months (hence why we still have a 1.2% uplift on last year compared to a 2% uplift when June’s figure was reported, in spite of a 1.5% drop between the two months) - this, again, is why Savills -2% for the year is the wrong figure for 2026 as their updated prediction states (unless you are in London of course).
The rental wrinkle deserves a paragraph of its own, because it cuts against the standard narrative. The average UK rent achieved in July was £1,837 per calendar month - against £1,849 in July 2025. Down, fractionally, year on year. Rental stock available sits at 323k against 319k a year ago, and new rental listings in July ran at 135.9k against 128.8k last July. I would not over-read one month of a mix-sensitive series - a heavy-London month or a student-cycle quirk can move this figure without anything real changing, so let me hedge it properly - that’s a complete reversal on June’s figures. But a rental market with more stock, more new listings and a flat-to-negative headline rent is not the "rents only go up" market of 2022-24, and it rhymes with what the affordability data has been whispering: the ONS Dataloft series this week put rent at 28.7% of gross income in June, down from 29.3% a year earlier (you never see any fanfare headlines with these sorts of bits of news, of course, because it doesn’t fit the mainstream narrative). Tenant affordability was always going to be the ceiling. It would seem to be doing ceiling-like things.
Chris - this is my weekly appreciation paragraph, with a double asterisk this week since I'm in the show as well as citing it. Thanks for what you do! If you want some help positioning yourself as a local market expert - as an estate agent or any form of property professional - give Chris a shout. Either way, give his channel www.youtube.com/@christopherwatkin a follow and some love, please - the Maidstone segment alone is worth it for anyone who still believes all agents are interchangeable, and if you are selling any property now or in the future, I’d recommend you watch it!
Dust off the Macroscope, then. A shortened data week by recent standards, but a well-shaped one: the PMIs, all of them, which finally split into two different economies; the July house price index from Lloyds, which is the index formerly known as Halifax, exactly as trailed in this slot last week; a public service productivity release from the ONS that lands 80 days before a Budget which will lean heavily on the very number it measures; and bringing up the rear, the contract states that we have to talk about the gilts and swaps - where, for once, the news is good for borrowers. Off we go.
Onto the PMIs first, my darlings of the real time economy - and this month the darlings turned up. Services: 52.1 on the final July reading, revised up from the 51.8 flash and a proper rebound from June's 48.8, back into expansion after two months below the line. New work rose for the first time in five months, with the survey crediting consumer spending, technology demand, and - I enjoyed this - hospitality boosted by the good weather and the World Cup. Manufacturing came in at 52.8, with output growth near a two-year high. The composite: 52.2, up from 49.3. So the economy-wide picture flipped from contraction to modest growth in a single month, which is either an inflection or survey noise, and the honest answer is that one month cannot tell you which. Chris’ Burnham Bounce after a paralysing squeeze waiting for Starmer to go? Not impossible. Two details stop me getting carried away. First, services employment fell for a TWENTY-SECOND consecutive month - and S&P's economists note that run now equals the joint-longest in thirty years of data collection, matching the 2008-09 crash and the dotcom aftermath. An expansion that will not hire is a strange sort of expansion, and I've been saying for the best part of a year that the jobs data is where the real story lives. Second, the inflation detail actually helped for once: input costs rose at the slowest pace since February, helped by fuel and gas - which is July's oil retreat showing up in the survey almost before it showed up in the barrel.
And then there is construction, which did not miss the memo about the party, but the repair was relative, not absolute. 44.7 in July, up from June's dismal 38.4 and well clear of the 40.0 consensus - the biggest one-month improvement in the recent record, and simultaneously the EIGHTEENTH consecutive month of contraction, the longest continuous decline since the global financial crisis. Housebuilding registered 41.8, its slowest rate of decline since October last year; commercial 46.8; civils a grim 38.3. So: the least-bad month since the spring, in the middle of the worst run since 2009. Is that a recovery? It is a bounce, certainly. Whether it is a live cat or a dead one with good hang time is precisely the question this week's Deep Dive was built to answer, with four independent sources who have no interest in flattering a diffusion index. Park it for now - the examination is coming.
Next up - the house price data, and a small piece of history: after 43 years, the Halifax House Price Index is no more. Same index, same methodology, new badge - it is the Lloyds House Price Index now, and as I said in this slot last week, Howard presumably gets no royalties. The rebranded series marked the occasion by going nowhere at all: the average UK price slipped £143 in July - which on £299,253 is a rounding error wearing a minus sign - and annual growth now stands at +0.1%, the slowest since November 2023. RBC's Anthony Codling called the market "suspended animation", and I can't improve on that as a two-word description of the headline. But the headline is a compression of two entirely different countries, and this is the bit worth your time: Northern Ireland +7.4% on the year, Scotland +3.6%, the North East +2.8%, the North West +2.1% - while the South East is down 2.0% and Greater London down 1.3% to £533,930. Nationwide's July read, covered here last week, told the same story from a different sample: +1.8% nationally, softening, with the south dragging. Two indices, £22k apart on the average price, in complete agreement on the shape. The national average is a fiction that nobody lives in; the north-south repricing is the actual market. For readers of this publication the practical read has not changed all year: the value zone remains the 2-3 bed terraces and semis in the unfashionable-but-employed towns, where yields still clear borrowing costs with room to spare - and where, not coincidentally, every regional index has prices grinding up rather than down. Suspended animation nationally; perfectly animated where we hunt. I'd also gently note that a 0.1% annual number, in a country running 2.6% CPI, is a 2.5% real-terms fall - the polite, orderly correction I have long thought more likely than the crash the doomers order every year and never receive.
Third sub-section, and an unusual one for this slot: public service productivity, quarterly, from the ONS on Friday. Bear with me, because this dry-sounding release is quietly one of the most Budget-relevant documents of the month. The headlines: total public service productivity rose 0.9% in 2025 after 0.7% in 2024 - and both years were just revised UP by 0.3 percentage points. Before anyone breaks out the bunting: the level remains 2.5% BELOW 2019. Healthcare - the biggest single component - grew 1.0% in 2025 and still sits 5.8% below its pre-pandemic level. Six years on, the state produces less than it did before Covid. Now, two wrinkles that the press release does not lead with. First, the upward revisions did not come from hospitals discovering efficiency; they came from the ONS changing how it measures healthcare inputs - a deflator rework on the spending that flows to GPs, dentists and ophthalmologists. Perfectly defensible statistics, but worth being clear that the improvement is partly in the measuring tape, and that’s rightly always approached sceptically. Second, and my favourite detail: the ONS rolling four-quarter measure says healthcare productivity rose 0.4% in the year to March 2026, while NHS England's own estimate for the same financial year says 3.4%. Same health service, same year, one measure eight times the other. Both are defensible on their own terms - different coverage, different methods, and the ONS itself says do not compare them directly - but when the two official answers to "is the NHS getting more productive?" differ by a factor of eight, I would suggest a little humility is in order from anyone quoting either with confidence. The confidence interval on the ONS's own 2025 figure runs from minus 1.3% to plus 3.2%, and these are, in their own label, statistics in development. Why does this matter to us, 80 days from a red box? Because the entire fiscal arithmetic - how much the Chancellor must raise from people like us - leans on assumptions about whether the public sector can produce more without costing more. If the honest answer is "we cannot currently measure it to within a country mile", then the Budget's productivity assumptions deserve to be read as hopes, not forecasts. And none of this is a dig at the people doing the work - it is a dig at flying a £1.3 trillion state on instruments this foggy.
Gilty, or not Gilty? Not Gilty this week, for a change - the first properly good week for borrowers since the spring. The closes: the 5-year gilt at 4.471%, down from 4.611% last Friday - 14 basis points of relief in five sessions. The 30-year at 5.670%, down from 5.787% - a shade under 12 points. Cast your mind back seven days: I wrote that 4.611% on the 5s "just looks too high" and that a sensible top end of the range was more like 4.5%. The market has, obligingly, gone and agreed with me inside a week - though I'd be the first to say that when a call comes good that fast, luck deserves at least half the credit, and the drivers were made in Houston and Washington rather than London: oil's round trip to the low $80s plus a negative US payrolls print equals lower yields everywhere, and gilts were dragged along in the happier direction for once. I also floated last week that the traders positioning for an uncertain August might end up disappointed while we borrowers ended up happy; one week in, so far so good, and I will resist the temptation to declare victory in week one of five. The swaps followed: the 5-year SONIA closed Thursday at 4.21%, against 4.02% a month ago and 3.62% a year ago. Sit with that last comparison - the 5-year swap is 59 basis points HIGHER than last August, with Bank Rate at 3.75% or higher the whole time and the yield curve showing the entire structure, one-year money out to the 30s, sitting a rung above where it sat a year ago. That is the imported-inflation era priced in sterling: the Bank has not moved, and your cost of money has. Two smaller items complete the picture. The Bank's market-implied path for Bank Rate one year out stood at 4.20% on the 4th of August, down from 4.30% on the 28th of July - so the market trimmed its hike pricing AFTER a 6-3 vote, which tells you how much work the oil price is doing in that spread (and the usual reminder: 4.20% is a market forecast, not the rate - the rate is 3.75% until at least the 17th of September). And the question I trailed last week - how much of the effective-rate climb has reached the shop window - got its answer on Friday: the Bank's quoted rate series on a 75% LTV two-year fix actually FELL a touch in July, to 4.79% from 4.81%, even as the effective rate on newly drawn mortgages rose to 4.36% in June. The shop window is holding its breath; the completions data is where the war premium lives. Moneyfacts' whole-market averages tell the same story standing still: 5.63% on the average two-year fix, 5.67% on the five-year, both essentially unmoved on the week and both a long way north of the sub-5% readings that opened 2026. Diary entries before we move on: the next MPC decision is the 17th of September - 39 days out - with the annual quantitative tightening envelope decided around the same meeting, the bond pile already down from its £895bn peak to £492bn, and my annual QT rant pre-booked and, on this week's evidence, likely to have fresh material. And the Budget briefing season is now live: between here and the 28th of October, every unfunded kite flown at a Sunday paper has a price in basis points. This week the basis points went our way. Enjoy it; I would not bet the farm on five more weeks of the same.
OK. Here endeth the lesson on current rates - here comes the Deep Dive.
Where are we going for the Deep Dive this week? Somewhere slightly unusual: all four sources point at the same building site. I flagged above that the construction PMI's bounce to 44.7 poses a question a diffusion index cannot answer on its own - and then the week's report flow did something helpful, which is to deliver four entirely independent witnesses. The NHBC published its Q2 registrations, which is the pipeline telling us what builders are committing to. The housebuilders and those who support them published half-year results in a cluster - Taylor Wimpey, Persimmon, Henry Boot alongside Ibstock and Travis Perkins - which is the coalface under oath, more or less. The HBF published "Payback for Good", its accounting of the Help to Buy scheme and its pitch for a successor, which is the demand-side fuel argument. And the Resolution Foundation published its Housing Outlook on whether the state can build what the Prime Minister has promised, which is the fourth wall of the site. So this week is a declared construction special - one question, four directions: the pipeline, the coalface, the fuel, and the state. Let's examine the cat for signs of life, Schrodinger-style observations, and anything else relevant:
Source one: NHBC new home registrations, Q2 2026 - the pipeline.
Theme one - the forward order book.
Summary: The National House Building Council, whose warranties cover the large majority of new homes built in the UK, reported 29,162 new homes registered in Q2 2026 - down 4% on Q2 2025. Registrations are the point at which a builder commits a plot to construction and purchases the warranty, typically weeks before ground is broken, making the series one of the earliest forward indicators of housebuilding activity available. The Q2 decline follows falls in previous quarters, with NHBC noting that developers are slowing housebuilding activity in response to market conditions.
The Propenomix Perspective: Registrations are as close as this industry gets to a revealed-preference survey - nobody buys a warranty for a home they doubt they'll build. So while the PMI was recording its least-bad month in over a year, the men and women signing the cheques were committing 4% fewer plots than in an already-weak 2025. I'd weight the cheque-signers. The gap between the two isn't a contradiction, mind: the PMI measures the direction of current activity, registrations measure conviction about next year. You can be slightly less miserable about today and still unwilling to commit to tomorrow - in fact I'd say that is precisely the industry's mood, and I hear it in every conversation I have with anyone holding or developing land at the moment.
Theme two - the completions paradox.
Summary: Against the registration decline, NHBC reported 32,973 new homes completed in Q2 2026, up 1% on the same quarter of 2025. Completions reflect decisions made 18 months to three years earlier, when the homes now finishing were started. The divergence - completions marginally rising while registrations fall - implies the industry is finishing more homes than it is starting, and that the gap between the two series will feed through to falling completions with a lag.
The Propenomix Perspective: This is the shape I'd draw on a whiteboard for anyone who thinks the housing supply problem is being solved: the past is still delivering while the future is being cancelled. Completions today are the echo of 2023-24 decisions; registrations today are the promise for 2027-28, and the promise is shrinking. It is the demographic pyramid problem transplanted into housebuilding - the pipeline is top-heavy with old decisions. For investors the read-through is unfashionable but I'll make it anyway: every quarter of registrations running below completions tightens the second-hand market three years out, particularly in exactly the family-house segment the volume builders have retreated from. Perhaps the most reliable forecast in this entire edition is that the mid-2020s under-building will be paid for, with interest, by the late-2020s buyer.
Theme three - what the counter actually counts.
Summary: NHBC's data substantially undercounts total UK housing delivery, since it captures only homes carrying NHBC warranties - roughly 70-80% of the new-build market historically, with coverage varying by year and excluding much of the self-build, conversion and some BTR output. The MHCLG's new-build Energy Performance Certificate series, an alternative leading indicator for net additions in England, recorded 202,667 new-build EPCs in the year to Q2 2026 on a duplicate-cleaned basis, with the 52-week rolling total at 209,771 in late July - slightly lower than the previous week.
The Propenomix Perspective: A short methodological sermon, because this week supplied a perfect cautionary tale: a widely-shared claim that Manchester built more homes than the whole of Greater London last year turned out, per James Gleeson's patient demolition, to rest on incomplete data. Housing supply statistics are a hall of mirrors - warranties, EPCs, completions, net additions, all counting slightly different things on slightly different lags - and anyone quoting one series as The Truth is selling something. My own preferred habit is triangulation: NHBC says the pipeline is shrinking, the EPC rolling total says delivery has plateaued around 210k and just started edging down, and the two together are more convincing than either alone. The direction is unanimous even where the levels disagree. That's usually the most honesty you can extract from this subject.
Theme four - synthesis: does the pipeline believe the bounce?
Summary: Taken together, the NHBC data describes an industry completing slightly more homes than a year ago while committing to 4% fewer, during the same quarter in which the construction PMI recorded its fastest improvement in the recent record while remaining in contraction. The registration series has not yet shown the stabilisation that the survey data hints at.
The Propenomix Perspective: So: witness one says the cat is dead, or at least not visibly breathing. Registrations are the purest forward bet in the dataset and they are still falling. I'd add one hedge in fairness to the optimists - registrations are lumpy, one quarter proves little, and Q3's figure will be collected in a quarter where mortgage pricing has at least stopped rising. If the bounce is real, registrations are where it must show up first, probably by the winter. Until then, I file the PMI's 44.7 under "less bad" rather than "better", and I suspect the builders would privately agree. Witness two, conveniently, is the builders themselves - under the mild oath of a results announcement.
Source two: the housebuilders' half-year reporting season - the coalface.
Taylor Wimpey, Persimmon, Ibstock, Henry Boot and Travis Perkins all reported within the fortnight; I'm treating the season as one composite source, since together they cover the volume builders, the land market and the materials chain.
Theme one - the sales rate, and the bulk-deal veil.
Summary: Taylor Wimpey reported a net private sales rate of 0.55 per outlet per week in the four weeks to 26 July, against 0.59 in the equivalent 2025 period - 0.53 excluding bulk deals, against 0.56. Cancellations ran at 18%, marginally better than last year's 19%. Persimmon reported net private sales up 6% at 0.72 per outlet per week in the five weeks since 30 June - but 0.59 excluding bulk deals, against 0.61 last year, and noted it continues to target growth in Build to Rent.
The Propenomix Perspective: Read those two statements side by side and the trick reveals itself: Persimmon's headline is UP 6% while its underlying open-market rate is DOWN - the difference is bulk deals, homes sold in blocks to institutions and housing associations rather than to families at full price. Taylor Wimpey's gap between headline and ex-bulk tells the same story more quietly. I'm not sneering at bulk sales - cash flow is cash flow, and in this market I'd take it too - but a sales rate propped up by wholesale channels is not evidence of returning retail demand, and retail demand is what the dead-cat question is actually about. On the coalface's own numbers, the private buyer is still scarcer than last year at every builder that publishes the split. Worth remembering when the sector's PR lands in September. Let’s be honest - those deals are usually not deals they want, they are deals that they have to take.
Theme two - the price of moving product.
Summary: Taylor Wimpey stated that underlying pricing has been broadly stable in recent weeks but remains on average approximately 2% below prior year levels. The wider results season pointed to continued use of incentives to support sales rates, against a backdrop in which the Lloyds index recorded new-build price premiums and overall market pricing essentially flat year on year.
The Propenomix Perspective: A 2% year-on-year decline in achieved pricing, delivered in the same market where the second-hand index reads +0.1%, tells you the discounting is happening at the margin where builders must move completed stock. And "broadly stable in recent weeks" is doing quiet work in that sentence - stable at a discount is not recovery, it is a market that has found the clearing price and dislikes it. The strategic point I'd pull out for our readership: when volume builders are achieving 2% below last year WITH incentives on top, the true like-for-like gap to the used market narrows further, and the second-hand buyer's premium for a warranty and a blank canvas is thinner than the headline premium suggests. I've said for years that patient buyers should price new-build against the builder's month-end rather than the brochure; nothing in this results season changes my mind, and the October year-ends are approaching. I remember 2023 well - housebuilders reporting an advance in new-build pricing of 2-3% in a year where the open market nominally fell around 2% - I said that 5% gap would come home to roost at some point, and indeed - we are there.
Theme three - the land market tells the truth.
Summary: Henry Boot's half-year trading update reported that a number of housebuilders have changed their land strategy, slowing acquisition activity, resulting in delays to transactions and increased use of deferred payment terms, and stated that given current market dynamics it expects 2026 plot sales to be materially below the prior year.
The Propenomix Perspective: If registrations are revealed preference, land buying is revealed preference with a decade-long consequence attached, and Henry Boot sits at exactly the till where that preference is expressed. "Slowing acquisition, deferred payments, materially below prior year" is the sector saying, with its land budget rather than its outlook statements, that it does not expect volume recovery on any near horizon. I find this the single most persuasive witness of the four themes, if I'm honest - land is the input you buy when you believe, and nobody is buying belief at the moment. The one counter-reading worth airing: a frozen land market also means land PRICES soften, and the builders who do step in over the next 18 months will be buying the margin that funds the next cycle's profits. Somebody always buys the bottom; the results season strongly suggests it hasn't been bought yet. For those of us further down the food chain, small-site land with consent is quietly becoming the most negotiable asset class in the country - a sentence I wouldn’t have dreamed of writing five years ago.
Theme four - synthesis: the materials chain and the margin.
Summary: Ibstock, the UK's largest brickmaker, reported a statutory loss before tax of £27 million against an £8 million profit in the prior period, citing a challenging backdrop, margin headwinds, and an impairment charge on mothballed facilities as market conditions delay recovery. Travis Perkins reported the UK construction sector remains subdued with depressed first-half activity, and noted that building materials price inflation remains hard to forecast given geopolitical and macroeconomic events.
The Propenomix Perspective: Mothballed brick kilns are the industrial version of falling registrations - capacity withdrawn because the order book doesn't justify the gas bill. So witness two agrees with witness one from five different boardrooms: activity subdued, private demand below last year, land appetite frozen, capacity coming out. Which sets up the awkward sequel that Travis Perkins is politely flagging: when demand does return, it will meet a supply chain that has shut kilns and shed capacity, into an energy market that a single Hormuz incident can reprice - and materials inflation will do what it did in 2021-22, at least in part. The bounce, if and when it becomes real, arrives with a cost problem pre-installed. I suspect the 2027-28 build cost conversation is being written in this results season, and almost nobody is reading it yet. For those who like coal-face reports - Noble Francis is well worth following on LinkedIn, he never misses a Brick Delivery report! Onwards - to the demand side, and the money.
Source three: HBF, "Payback for Good" - the fuel.
Theme one - the £1.74bn arithmetic.
Summary: The Home Builders Federation's report, published this week, draws on Homes England figures to account for the Help to Buy equity loan scheme, which ran from 2013 to 2022 and assisted 387,278 households to purchase a new-build home, 328,346 of them first-time buyers. Of the loans granted, 213,713 - around 55% - have been fully repaid. Closed loan accounts had an origination value of £11.98 billion against a repayment value of £13.22 billion, a £1.24 billion surplus representing a 10.4% return. Interest payments have generated a further £510 million for Homes England since 2018/19, including £151.9 million in 2025/26, giving a combined Exchequer benefit the HBF estimates at around £1.74 billion of "profit".
The Propenomix Perspective: Credit where due: as a piece of accounting, this is more honest than most lobbying documents, and the core fact - the taxpayer has so far made money on the loan book - is real and under-reported. But let me be careful about what the £1.74bn is and isn't. It is the return on the loans that have CLOSED - disproportionately the ones where prices rose and owners could afford to repay. The 45% still outstanding includes the harder tail, and equity loans mark to market both ways. And a 10.4% total return over up to thirteen years is, whisper it, a fairly limp lettuce next to the government's own cost of borrowing across the period. So: profitable, yes; a triumph of public investment, not quite. The truthful framing is that the scheme roughly washed its face in cash terms - which, for a demand subsidy, is genuinely rarer than it should be.
Theme two - the supply claim.
Summary: The report argues Help to Buy underpinned a doubling of housing supply, noting that net additions reached a peacetime low of 124,000 in 2012/13 (136,000 gross), five years after the crash, and that annualised planning approvals rose from a consistent 150,000-200,000 range to above 250,000 by late 2015, peaking above 330,000. It attributes the recovery to the forward visibility the scheme gave builders for land acquisition and investment in labour and skills, and notes that consents and investment in new sites have declined consistently in the four years since withdrawal - the first period in 60 years with no government support scheme for buyers in place.
The Propenomix Perspective: The correlation is real; the attribution is where I'd haggle. Supply doubled from a 2013 trough that coincided with Help to Buy, yes - but also with a near-zero Bank Rate, Funding for Lending, a recovering economy and the base effect of starting from the worst year since the war. Untangling those is properly hard, and the HBF, understandably, does not try very hard to untangle them. What I will grant them without argument is the forward-visibility mechanism: builders build what they can pre-sell, and a state-backed deposit for a third of buyers is a formidable pre-sales machine. The four-year slide in consents since withdrawal is at least consistent with their story. My honest position, held for years: Help to Buy raised output AND prices - the question was always the ratio, and the ratio was never as favourable as the sector claimed nor as damning as the critics did.
Theme three - the inflation question, and the missing evaluation.
Summary: The report includes a section directly addressing what it terms the "inflationary" arguments against the scheme, contesting research that found Help to Buy raised new-build prices, particularly in areas with constrained supply. Separately, the HBF notes that the government completed its own formal evaluation of Help to Buy earlier this year; the evaluation has been presented to ministers but not published, despite repeated calls from the federation for its release.
The Propenomix Perspective: Two observations, one per hand. On the first: a trade body whose members sold the product is not the referee I'd choose for the inflation question - the serious academic work found meaningful price effects precisely where supply couldn't respond, which is exactly where you'd predict a demand subsidy to leak into price rather than volume. Sowell's first lesson of economics, from the top of this edition, was written for schemes like this. On the second hand, though: the HBF is dead right about the evaluation, and I'd go further than they politely do. A completed, taxpayer-funded evaluation of a £29bn-lifetime intervention sitting unpublished in a ministerial drawer, while the government weighs whether to launch a successor, is not a neutral act - whatever it says. Publish it. If it vindicates the scheme, the HBF gets its evidence; if it doesn't, the taxpayer gets fair warning. Either way we stop arguing in the dark, which suits nobody except whoever benefits from the dark. The Prince of Darkness - he liked the dark, after all (Too cynical? On this file, I doubt it.)
Theme four - synthesis: the successor proposal.
Summary: The HBF proposes a replacement equity loan scheme for first-time buyers on new-build homes, structured around a 20% equity loan and part-funded by a financial contribution from home builders themselves, arguing this would reduce the cost and risk to the taxpayer, restore forward visibility to the industry, and address what it describes as a dearth of affordable mortgage lending for buyers with small deposits.
The Propenomix Perspective: The developer-contribution element is the one properly new idea here, and it deserves to be taken seriously rather than waved through or laughed off. Seriously in favour: it partially aligns incentives - if builders co-fund the subsidy, they co-own the risk, which blunts the purest version of the "public money, private landbank gains" critique. Seriously against: a contribution funded by builders is, one way or another, priced into the homes, so the buyer part-funds their own subsidy through the sticker - Sowell's second lesson, politics disregarding the first, wearing a hard hat. Would it move the pipeline? Probably, yes - the pre-sales machine worked before and would work again, and this week's other three sources all testify to how badly the pipeline wants fuel. Would some of it leak into price? Also probably yes. Eighty days from a Budget, with a Chancellor needing a housebuilding story that doesn't cost headline billions, I'd rate the odds of some version of this appearing rather higher than the commentary currently does. File under: watch the kite-flying.
Source four: Resolution Foundation, Housing Outlook Q3 2026 - the state.
Theme one - the £39 billion ring-fence arithmetic.
Summary: The Resolution Foundation's quarterly Housing Outlook examines how the Prime Minister's council housebuilding ambition could be delivered through the government's £39 billion, ten-year affordable homes grant programme. Its central estimate: if the entire programme were ring-fenced for social rent - the deepest-subsidy tenure - just under 25,000 affordable homes could be delivered annually, against the government's stated target of 30,000 social homes a year.
The Propenomix Perspective: Run the division and you see the problem immediately: £39bn over ten years is £3.9bn a year, and at the grant rates per social-rent home in the £150k-plus territory that prevail across much of the country, £3.9bn simply does not divide into 30,000 - it divides into something with a two-handle. That is the entire report in one long-division sum, and I admire the Foundation for doing arithmetic where others do adjectives. The government's escape routes are limited: more grant (the Budget has no room), cheaper tenures (which changes the promise), cross-subsidy from councils' own borrowing (capacity varies wildly, with many overleveraged already), or quietly redefining "social". I'd guess at a blend of the third and fourth, if I'm honest. The ambition is real and, I'd say, sincere - the arithmetic is just bigger than the cheque.
Theme two - the tenure trade-off.
Summary: The gap the Foundation identifies stems from tenure economics: social rent homes, let at roughly half market rents, require substantially more grant per home than shared ownership or affordable rent, so every social-rent home delivered consumes the budget faster. A programme optimised for headline unit numbers would favour shallower-subsidy tenures; a programme optimised for the households in deepest need favours social rent and delivers fewer homes. The report frames this as the central, unavoidable choice within a fixed envelope.
The Propenomix Perspective: This is the most Sowell-shaped paragraph of the week: scarcity means the choice between MORE homes and DEEPER help cannot be wished away, only made - and the first lesson of politics is to announce both and choose neither. For our readers there is a practical angle hiding in the tenure maths: whichever way the blend lands, the private rented sector remains the overflow vessel for everyone the programme cannot reach, which on these numbers is most of the waiting list for most of the decade. The 119-year social housing queue we examined in June is not going to be materially shortened by 25,000 homes a year against a million-household list; I take no pleasure in that sentence, but the arithmetic wrote it, not me. Landlords providing decent stock at the affordable end are, whether the discourse likes it or not, load-bearing for another decade. Behave accordingly - the responsibility runs both ways.
Theme three - the Burnham factor.
Summary: The report lands three weeks into Andy Burnham's premiership, which opened with a declared national drive to end rough sleeping at the earliest opportunity, and days after a devolution programme granting mayors a share of income tax revenues - covered in last week's Deep Dive. The council housebuilding ambition sits alongside these as the new government's housing platform, with delivery mechanisms and departmental budgets yet to be tested at a fiscal event; next week's statutory homelessness statistics will provide the baseline against which the rough sleeping pledge is measured.
The Propenomix Perspective: I'll keep the politics short since we did the constitutional piece last week, but the connective tissue matters: a Prime Minister who made his name on Manchester's housing-first record (of which there are varying analytical reports, some sympathetic and some not - but too political to get into without an individual deep dive) now owns the national version of the problem, with a Chancellor who has 80 days to fund it and a grant programme that - per theme one - is already fully subscribed by its own promises. Something has to give, and my suspicion is that the giving happens in the definition of terms rather than the size of cheques; watch for "affordable" doing a lot of heavy lifting on the 28th of October. That said - and here's my hedge - Burnham is the first PM in a generation who has actually run housing delivery at scale, and machine-knowledge occasionally beats money. I'm prepared to be pleasantly surprised. I'm just not pricing it in.
Theme four - synthesis: the verdict on the cat.
Summary: Across the four sources: registrations fell 4% while completions rose 1%, indicating a shrinking forward pipeline; the volume builders reported private sales rates below last year once bulk deals are excluded, pricing 2% down, land acquisition slowing and materials capacity mothballed; the industry's demand-side proposal awaits a Budget decision; and the state's programme, fully ring-fenced, delivers below its own target. Build-to-rent starts, per RE:UK's Q2 report, fell 79% year on year to 3,455 - a further contraction in the one pipeline that had been growing. The construction PMI's July improvement to 44.7 is, at the time of writing, uncorroborated by any hard forward indicator.
The Propenomix Perspective: Verdict time, then, and I'll show my working like the honest witnesses did. For the bounce being real: the PMI's improvement was broad, mortgage pricing has stopped rising, oil just handed back a fifth of its premium, and sentiment often turns before the data that vindicates it. Against: every forward commitment measure - registrations, land deals, kiln capacity, BTR starts - is still pointing down, and the people spending actual money are behaving as though the winter will be long. My call, hedged as it deserves: this is a dead cat with unusually good hang time - a genuine deceleration of the decline, which matters, but not yet a turn. What would change my mind, so you can hold me to it: two more construction PMI prints above 45 with housebuilding above 45, a positive quarter of NHBC registrations, or a Budget demand-side surprise on the 28th of October. Any one of those and I'll reopen the file, cheerfully. Until then the strategic conclusion stands where it has stood all year: the supply squeeze of the late 2020s is being manufactured right now, in mothballed kilns and unsigned land deals, and the patient buyer of ordinary houses in ordinary towns is on the right side of it.
The Week Ahead. The diary earns its keep next week, because Thursday is a monster. Wednesday the 12th warms up gently with MHCLG's energy efficiency statistics on new dwellings. Then Thursday the 13th delivers six at once: statutory homelessness figures for England, which become the baseline for the new Prime Minister's rough sleeping pledge and reconnect to our 119-year queue thread; the monthly GDP estimate for June, closing out the second quarter; official construction output, which gets to agree or argue with the PMI's bounce; the MoJ's mortgage and landlord possession statistics alongside UK Finance's arrears and possessions data - the pair I watch for any crack in the "quietly reassuring" household credit picture, and so far there hasn't been one; and RICS's July Residential Market Survey for the sentiment read. Monday the 17th brings Rightmove's asking price index, Tuesday the 18th the ONS labour market overview - where, as ever, we'll do employment, unemployment AND inactivity, because the unemployment rate alone is the least informative number in British statistics - and Wednesday the 19th the July CPI print plus the ONS private rents and house prices release, the inflation pair that will frame the September MPC conversation. Friday the 21st closes with public sector finances, feeding the Budget countdown. Across the water, the US inflation print lands midweek as the first test of whether the payrolls shock was a blip or a turn. And the standing reminder of the season: the Sunday papers are market-moving events between now and late October - we'll sort the kite-flying from the signal right here. Set your alerts accordingly.
As we get towards the end for this week - I genuinely can't wait for the next workshop. Thursday the 1st of October, in Manchester, with Rod Turner: joint ventures and M&A, which on this week's evidence is exactly the muscle the next few years will reward - the deals, partnerships and acquisitions that get done while solo capital waits for certainty that never comes. The Super Early Bird discount is running at better than 20% off and the VIP dinner on the night is the best value in the room, as the regulars will tell you. Book your tickets: tinyurl.com/pbwoct26
Above all - please remember to Keep Calm, ALWAYS listen to or read the Supplement, and Carry On. Look at the week we've just had: oil round-tripped a fifth, the 5-year gilt handed borrowers 14 basis points, a 43-year-old index changed its name, and a construction sector had its least-bad month in a year and a half - and through all of it the fundamentals we track every week did not move an inch. Too few of the right homes in the right places, with this week's evidence saying fewer still are coming; a sales market transacting at decade-average volumes that pays proper prices to the correctly priced and shows the door to everyone else; rents finally testing the affordability ceiling; yields that still clear the cost of money with room to spare for the well-bought. Eighty days to the red box, thirty-nine to the MPC, and the dangerous months will do what they always do to the impatient. We are not the impatient. KCCO!