Nisha WijayaChief Operating Officer · Essays & perspective
← All essays

Construction delivery5 min read

A contractor fails in its payables before it fails on site

Owners of large construction programmes watch the programme, the quality records and the safety statistics. The earliest sign that a main contractor is failing usually appears further down, among the subcontractors and suppliers it has quietly stopped paying.

Hai Phong, September 2022. The welding crews on the roofs of the third phase of our park had thinned from about forty people to eleven in a fortnight. The main contractor's weekly report called it a regional labour shortage and showed the programme slipping by nine days. Our site manager did something the report could not. He sat in the workers' canteen at lunchtime and asked the welders when they had last been paid. Most of them said seven weeks.

Ten weeks later the contractor stopped work on all eighteen factories in the phase, and we spent the following year finishing them ourselves. I describe that year elsewhere on this site. Here I want to make a narrower argument, which I have come to hold after watching three contractors fail on projects I was responsible for. A main contractor fails in its payables long before it fails on site, and owners who wait for the programme to tell them are choosing to learn late.

A quiet yard in East Java

My first lesson came in 1998, when I was a site engineer for a Surabaya contractor building a garment factory south of the city. The rupiah had lost most of its value in a few months. Our steel fabricator had borrowed in dollars and was going under without telling anyone. What I noticed was the silence: one Monday morning the grinding and hammering from the fabrication yard had stopped. The fabricator's men had not been paid for six weeks and had gone home. Our programme, issued the previous Friday, showed steel erection on schedule.

The programme was not dishonest. It recorded what the planners believed, and planners believe what the next layer down tells them. A construction programme is a forecast built on the assumption that everybody in the chain will keep turning up. The payment chain is the evidence of whether they will. A contractor running short of cash does not stop paying everyone at once. It stretches the people least able to complain: labour subcontractors, small fabricators, plant-hire firms and workers paid in cash at the end of the month. They keep working for a few weeks out of habit and hope. Then they leave, and the programme catches up with what the payables already knew.

Three questions at every monthly valuation

Every owner of a construction programme certifies a monthly valuation and pays its main contractor against it. That monthly moment is the cheapest place to look further down the chain, and since 2023 our project teams in Vietnam and Indonesia ask three questions at it.

The first is whether the contractor has paid its subcontractors and main suppliers for the work we paid it for last month. We ask the contractor to declare the position, and then we ask a rotating sample of subcontractors directly, in writing and in person on site. The answers rarely match exactly, and a gap of a few days is normal. A pattern of 30, 45 or 60 days, concentrated on the smallest firms, is a warning.

The second is whether the labour on site matches the labour the programme needs. We count people at the gate every day, by trade, and compare the count with the contractor's own resource plan. A falling headcount in a critical trade, with no change to the plan, is the same signal I heard in 1998, only measured.

The third is whether anyone in the chain has made a formal claim, sent a demand letter or held back a delivery. Suppliers of concrete, steel and plant usually know before anyone else, because they are the first to move a struggling contractor onto cash terms. Our procurement team stays in touch with the main suppliers to each park and tells me when a contractor working for us has been put on cash on delivery.

None of this requires an owner to manage the contractor's business. It requires an owner to read evidence that is already lying on its own site. On contracts above US$20 million we go one step further and pay through a project account, from which subcontractors on an agreed list are paid directly at each valuation, with the contractor's consent written into the contract. Contractors price that arrangement, modestly, and a few decline to bid. I accept both effects.

Why owners prefer not to ask

When I put this argument to other owners I hear three objections, and each has something in it.

The first is that the payment chain is the contractor's affair, and an owner who reaches into it blurs the contract. That is partly true. We do not instruct subcontractors, and we pay them directly only where the contract allows it and on legal advice. But asking a subcontractor whether it has been paid is not instructing it. An owner is entitled to know whether the people building its asset will still be on site next month.

The second objection is cost. Gate counts, sample letters and supplier conversations take time from project teams that already have plenty to do. On our programmes the cost is roughly one person per large project, which is small beside a year of delay.

The third objection is rarely spoken: owners prefer the contractor's version because it lets them carry a fixed-price contract on their books as if the risk had gone. A fixed price transfers risk only for as long as the contractor can bear it. When it cannot, the risk returns to the owner, usually larger and at the worst moment. In Hai Phong it returned when eleven of the eighteen factories were already leased to manufacturers with production plans of their own.

I do not claim that watching payables would have saved that contractor. Vietnam's property and corporate-bond markets tightened sharply in late 2022, and a contractor exposed to residential developers had problems no single client could solve. What it would have given us was time. We first heard about unpaid welders in August and terminated in December. With the three questions in place, I believe we would have seen the pattern by June and prepared to step in while the contractor's own staff were still willing to hand over drawings and site records. Those four or five months account for most of the delay our tenants suffered.

For directors the practical point is short. When management reports on a large construction programme, ask what it knows about payment further down the chain, how it knows and when it last checked. If the answer is the contractor's monthly report, the board is hearing what the contractor's planners believe. The welders usually know sooner.