The latest MPLC market brief & model read
Updated every Monday & Friday · Week in review · 2026-09-04 · Page 1 market brief · Page 2 what our model is telling us. The market-brief page is an automated research scan; the model read is generated from our data. Information & forecasts only.
The week's real story played out in the bond market rather than any boardroom. Australia's 10-year government bond yield increased to 5.19%, the highest since July 2011, as a overshot domestic inflation print in July, when headline inflation rose to 3.6%, ahead of expectations of 3.3%, collided with the return of strikes between the US and Iran and a wave of record corporate debt issuance, including Australian dollar offerings by Alphabet. NAB has now become the first of the major banks to forecast a September move, pencilling in a 0.25 percentage point rise to 4.6%, with the decision due at the RBA's meeting on 28-29 September 2026. For property, this matters a lot: our model had the 10-year sitting around 4.92% as at late August, so a jump past 5.19% this week widens the gap between where borrowing costs sit and where the all-property market is pricing (a 5.55% cap rate against a 6.61% fair-value benchmark), adding fresh pressure for cap rates to drift wider from here.
Buyers, though, aren't waiting for clarity. Half-year transaction data confirms Australian commercial property transactions reached $19 billion in the first half of 2026, a 16 per cent increase on the same period last year, with domestic institutional investors stepping in as offshore purchasing retreated, with offshore buying now just 21% of the total sales volume. The headline deals told the story: Sydney-based Goodman Group's $2.65 billion industrial deal with Washington H. Soul Pattinson underscored why industrial remains the cheapest sector on our numbers, while office finally found a buyer of scale in 100 Mount Street in North Sydney, which sold for $558 million to Investa, BGO and Cliffbrook Capital — a rare vote of confidence in a sector still priced well above our modelled fair value.
Underneath it all, the AI infrastructure build keeps quietly reshaping the alternatives space: CommBank estimates data centres will add around six percentage points to real business investment growth in 2026. And landlords with defensive income held up nicely too — Charter Hall Retail reported record portfolio occupancy of 99.1% as of June 30, 2026 and a net-lease book priced at a weighted average cap rate of 5.05%. The week leaves things poised on a knife-edge: buying momentum is real, but if bonds keep climbing into that September 29 RBA decision, cap rates across every sector will feel the pull.
In plain terms: right now the model likes Industrial best on value, and is most wary of Retail. Looking five years out, it expects Industrial to earn the most — about 8.7% a year — while Retail trails at roughly 6.7%. Why are property values under pressure? Because the ten-year government bond now pays around 4.9%. When something as safe as government debt pays that much, a building has to offer more to tempt a buyer — and that quietly drags values down. What's holding them up is the steady wave of overseas money still buying in — the one signal we've actually proven moves prices. The table below is that same picture, in numbers.
| Sector | Cap rate now | Fair value | 5-yr return | Worst case |
|---|---|---|---|---|
| Office | 6.32% | 7.05% | 7.8% | -17% |
| Retail | 5.59% | 6.78% | 6.7% | -19% |
| Industrial | 5.40% | 5.99% | 8.7% | -11% |
| All Property | 5.55% | 6.61% | 7.3% | -17% |
| Alternatives | 5.87% | 6.61% | 8.2% | -15% |
A quick guide: "cap rate" is the yearly rent as a share of the price — like an interest rate on a building; a higher number means cheaper. "Fair value" is what our model says that rate should be given today's interest rates. "Worst case" is a bad-but-realistic five-year fall in value.
Cheapest to priciest, the order runs Industrial then Alternatives then Office then All Property then Retail. We hunt for value, not hype — the best long-run buys tend to be where the crowd has already sold off.
Independent research — we don't develop, invest in or broker property, and every figure is built from public data. How it's built →