An idle line costs less than a price you cannot take back

There is a sentence I have heard in nearly every capacity discussion of my career, in a joint-venture boardroom in Wuhan, at a parcel carrier in Hangzhou and now in Changsha: a plant must run. It is said as though it were a law of engineering. It is really an accounting habit, and in the price war that has run through China's electric commercial-vehicle market since early 2024 it has cost the industry a great deal of money.

The logic is familiar. A vehicle plant carries heavy fixed costs: depreciation on the body shop and paint shop, the salaries of engineers and supervisors, the energy needed to hold a paint shop at temperature whether it paints ten vans or a thousand. Every van not built leaves part of that cost unabsorbed. So when a competitor cuts its price, the reflex is to cut too, keep the volume and protect absorption. Each company's reflex is reasonable. Together they produce an industry with more capacity than buyers and prices falling faster than costs.

Why a full plant feels safer than it is

The first problem is that unabsorbed fixed cost is visible and lost margin is not. When a line stands idle, cost per vehicle rises in the monthly report, the plant director has to explain it and the local government asks why the line it helped to build has gone quiet. When a price is cut, the loss is spread thinly across every van sold, for as long as the lower price lasts, and it appears as a slightly weaker gross margin that can be blamed on the market.

At Tanzhou in March 2024 the arithmetic was stark. Two competitors had cut the list price of their 3.5-tonne vans by about 12 per cent. Matching them would have taken roughly RMB 17,000 off each core van, more than RMB 1.5 billion of gross margin over a year at the volume we expected. Idling the second assembly line at our Chongqing plant would cost about RMB 48 million a year in depreciation, preservation and the wages of people we redeployed rather than released. The gap between those two figures is the whole argument. Yet in the meeting where we discussed it, the idle line felt like the larger risk, because it was the one you could photograph.

The second problem is that a full plant hides whether demand is real. Our dealers and fleet customers were holding 2.6 months of our stock that spring, more than twice the normal level. Running both plants on two shifts would have added to that stock, and our reports would have shown strong production and healthy utilisation for two or three quarters. Then the stock would have had to be cleared, probably at a further discount. A plant running flat out into a full channel is not efficient. It is borrowing from next year's prices.

What a cut list price does to the next three years

A list price in our market is hard to raise once it has been lowered. Fleet buyers are professionals who remember what they paid, and parcel franchisees compare notes in group chats within hours. Because we offer price protection to customers who commit to their orders early, a cut also reaches back into vans already allocated. In 2024 that cost us RMB 52 million on the models where we did cut, even though the core van's price held.

The damage outlasts the price war itself. Lower margins reduce what a manufacturer can spend on battery development, service networks and quality, which are what a fleet customer values most over a van's eight-year working life. Suppliers feel it next. When carmakers cut prices they ask suppliers for reductions, and a supplier under pressure economises in places the carmaker cannot see. I spent enough years in supplier quality to know where that ends: in a moulding batch that nobody had time to check properly.

There is also a policy direction that operators should not ignore. Since 2024 China's leadership and its industry ministries have spoken repeatedly against disorderly, destructive competition, which commentators call involution, and have urged carmakers to stop selling below cost and to pay suppliers promptly. A company whose recovery plan depends on outlasting its rivals through price cuts is betting against a direction the authorities have stated plainly.

Idling a line without abandoning a town

None of this makes idling capacity easy, and I do not want to understate what it costs people. When we idled the Chongqing line in July 2024, 640 people worked on it. We moved 410 to other work: the pack line at Changsha, a second shift on Chongqing's remaining line and our service network, which was short of technicians. 230 chose voluntary severance, paid above the statutory minimum. Some of them had moved their families to Chongqing two years earlier because we had asked them to. I met as many as I could, and those conversations were among the hardest of my working life.

The district government that provided the plant's land was unhappy and withheld an incentive instalment. I made a mistake there. It learnt of the decision from our announcement, and I should have gone to see it before the board paper was written. A local government that has invested in a plant has a legitimate interest in it, and its trust is built over decades, not quarters.

Three practices made the decision workable. First, we preserved the line rather than dismantling it, so that the choice could be reversed when demand returned; I underspent on that, and the cost of restarting has risen. Second, we divided demand in our capacity plan into three layers: orders customers have contracted, orders the channel can absorb without stock rising, and orders that only a price cut would create. The board sees all three, and only the first two may justify a shift pattern. Third, we competed on what a fleet buyer measures over the life of a van: an eight-year battery warranty and repairs made in the customer's own depot.

The outcome was not a triumph. Our market share fell from 9.8 to 8.6 per cent in 2024, and some customers left us for cheaper vans. But gross margin held at 13.9 per cent, we stayed profitable while larger rivals did not, and by 2025 deliveries had recovered to 138,000 with our Changsha plant at 92 per cent utilisation.

The lesson I would offer other operating chiefs is a narrow one. Before approving a shift pattern, ask which layer of demand it serves. If the answer is the layer that exists only because of a price cut, the plant is not running for its customers. It is running for the monthly report, and the report will not pay for the next three years.