A fixed price from a builder is not a transfer of risk

For most of my career the standard advice to an owner commissioning a building was to agree a fixed price with a design-and-construct contractor and push as much risk as possible across the table. Lenders liked it, boards liked it and investment committees could approve a single number. I no longer think it is the right default for complex buildings. A fixed price is only as good as the builder's capacity to absorb the risks it has accepted. For apartment towers and other complex projects, owners should keep the risks they can carry better than the builder, pay openly for the builder's work and stop paying a premium for certainty that may not exist on the day it is needed.

I should be clear about what this argument is not. It is not a claim that builders are untrustworthy, or that fixed prices never work. Corvalla still uses them for warehouses, where the design repeats, the ground is well understood and the trades are few. It is an argument about which party should hold which risk, made by an owner who was on the wrong side of the question for years.

What the lump sum was actually buying

When we tendered a 410-home rental tower in Parramatta in 2022, six builders were invited to bid a fixed price. Four declined. The two that bid came in 17% and 18% above our cost plan, and both excluded movements in steel and concrete prices, the risks we were most worried about. One builder's managing director told me he would not sign another fixed price for a tower of that size at any margin, because his last two had nearly broken the company. He was being candid, and the wider evidence supported him: several large and medium builders had collapsed around the country in the preceding year, many of them holding fixed-price contracts that had turned into losses as costs rose.

That conversation made me ask what we had been buying with lump sums. In a stable market, a fixed price buys a builder's judgement about risk, priced with a margin for being wrong. In an unstable one it buys something else. If the builder is well capitalised, it buys a large premium. If it is not, it buys a promise that holds until the losses exceed the builder's equity, after which the owner inherits an unfinished building, unpaid trades and a replacement contractor who will price every risk again. The owner never transferred the risk. It lent it to the builder, on terms that are repaid only when things go well.

Keep the risks the owner can carry

The principle we adopted is old and easy to state: give each risk to the party best able to manage or bear it. What changed was our willingness to apply it to ourselves. Some risks an owner holds better than a builder. Ground conditions are one, because we bought the site and commissioned the investigations. Design development is another while we are still making choices about the product, as we are in rental housing, where layouts, amenities and building services shape operating costs for decades. Commodity price movements over a three-year programme are a third, because a group with a balance sheet of several billion dollars can absorb them more cheaply than a builder whose equity is a few per cent of its turnover.

Other risks belong with the builder: productivity, programme, coordination of trades, safety and the quality of the work. Those are what a builder's margin should reward and what its performance should be measured against.

At Parramatta we turned that division into a two-stage contract with Ferngate Construction. In the first stage, which we paid for, Ferngate completed the design with our consultants and tendered the main trade packages on an open book. In the second, we agreed a price for the works with a fixed builder's margin of 4.5%, shared savings and overruns equally within a 5% band, and kept the ground risk and the price movements on four major packages, with our joint exposure to those movements capped at A$14 million. The agreed price was A$321 million, against fixed bids of A$349 million and A$352 million that had still excluded price movements. The final cost was A$327 million.

The price of an open book

This approach is not free, and owners who adopt it should say so to their boards. We raised development contingency from 5% to 8%, because we were now holding risks we had previously pretended to pass on. Our project team needed cost managers who could read a builder's trade accounts, and we hired them later than we should have. Our board received quarterly risk reports and asked harder questions than it had for any fixed-price building. Our capital partner, a pension investor, initially preferred the higher fixed bid and accepted the two-stage contract only with a cap on our exposure and monthly reporting. Each of those conditions was reasonable, and each cost us time.

We also learnt where our first attempt was wrong. Most of the overrun came from the facade, which we left inside the shared band although its design was the least resolved of any package at the end of stage one. A risk the owner has not finished defining should not be shared as if it had been. On our next tower we priced the facade separately, with its own contingency.

The benefit that surprised me most appeared in the market for builders. When we tendered our next housing tower in 2024 on the two-stage model, five builders bid rather than two. Better-capitalised firms that had stopped pursuing fixed-price apartment work were prepared to compete again, and the conversations during the tender were about how to build rather than about exclusions.

I would ask any owner's board to take its development pipeline and, for each project, list the risks the contract claims to transfer and what would happen to each one if the builder failed halfway through. Where the honest answer is that the risk comes back to the owner, the owner should consider holding it from the start, paying for it openly and measuring the builder on what the builder actually controls. The number in a fixed-price contract is comforting. It is worth only what stands behind it.