China's Record 22% Auto Collapse Is a Subsidy-Formula Failure — and 8.2 Million Exports Now Set the Floor Under US Prices

On October 26 the China Passenger Car Association cut its 2026 retail forecast to 18.5 million vehicles, 22% below the 23.74 million peak of 2025 and below the 19.29 million recorded in 2020, ending a five-year growth streak with the steepest annual decline the market has ever posted. The association's secretary general, Cui Dongshu, now expects only about 2% growth in 2027, to roughly 18.87 million units — which would still leave the market a fifth below 2025.

Three details separate this from an ordinary cyclical dip. First, the base is artificially strong: buyers pulled purchases forward into 2025 ahead of the January purchase-tax change, so 2026 is being compared against a pulled-forward peak. Second, the damage is concentrated in the electrification that policy was supposed to champion — new-energy sales fell 16.7% year over year in September even as their share of the market hit a record 67%, while combustion cars fell 43%. Third, the industry entered the downturn with slack: capacity utilization was 73.2% in 2025, below the 75% line described as healthy, so fixed costs cannot be absorbed by the home market at all.

Here is the arithmetic that broke the market. Until 2026, China's scrappage and trade-in programmes paid fixed sums of cash. For 2026 the State Council rewrote them as a percentage of the new car's price, with caps left unchanged, and the subsidy pool now shrinks precisely as the buyer gets cheaper. A RMB 70,000 new-energy buyer now collects RMB 8,400 where the old fixed rule paid RMB 20,000 — a 58% cut for the exact customer the programme was designed to put behind the wheel.

The 2027 rebound case is built on cost relief rather than demand repair: the CPCA expects about 2% growth to roughly 18.87 million units, helped by lower battery material costs, more competitive models and destocking. That is a market still roughly one-fifth below 2025, and it assumes the subsidy formula stays as written. S&P Global's January 2026 auto brief reached a harder conclusion: extending subsidies was unlikely to stop the 2026 sales drop.

Capacity that has no home goes where it can. CAAM data show 8.21 million vehicles shipped in the first nine months of 2026, up 65.8% year over year, while domestic sales fell 22.4%. Exports are now running above one million units a month and approaching 40% of Chinese vehicle shipments. One in every four cars built in China is now sold abroad, which is why the home-market collapse is showing up as price deflation in other people's markets rather than as a Chinese earnings crisis alone.

Here is the part the headline hides. A US tariff of 100% on Chinese electric vehicles means the cheapest cars in the world never appear on American price lists, so the deflation does not show up as cheaper cars in showrooms. It shows up as a floor under prices everywhere else, and as a pipeline aimed at the US market through Mexico and Canada. American price lists are defended by a wall, not by a moat — and the wall is only as strong as the enforcement around it.

The route that matters is localization, and it is already being priced. Nissan warned in September that Chinese carmakers will localize in North America, a pointed remark from the owner of the Mexican plant now in the bidding. Mexico has held its 50% wall on Chinese-built cars while asking Washington to cut the 25% Section 232 rate on Mexican-built cars to 15% — a request that, if granted, hands the cheapest Chinese capacity a duty-free path into the United States.

The upstream layer is where the Chinese system has already won a beachhead. Ford's Michigan plant started building the first US-made lithium-iron-phosphate cells in June 2026 using CATL's licensed chemistry, with launch-ready cells targeted before the end of the year. The cost curve of Chinese battery technology is now embedded in American-built vehicles, which is why the Chinese answer to a price floor is not a tariff negotiation but a technology licence.

The earnings split is already in the filings. BYD reported first-half revenue of RMB 344.8 billion, down 7.1%, with net profit down 21% to RMB 12.3 billion, before third-quarter revenue rose 13.6% to RMB 221.5 billion and net profit reached RMB 10 billion — the first sign that exports are replacing the domestic hole. SAIC Motor shows the other side: first-half operating revenue of RMB 298.7 billion, down 0.3%, with net profit attributable to shareholders of RMB 5.15 billion, down 14.4%, and nine-month revenue of RMB 455.2 billion against RMB 469.0 billion a year earlier. Flat revenue and a 30% fall in nine-month profit is what a demand collapse looks like before the inventory clears.

The facts are narrow and hard. China's retail market is 22% smaller in 2026 on the CPCA's own forecast. The 2026 subsidy pays a percentage of the purchase price under a cap that was not raised, which pays the expensive buyer and starves the cheap one. Exports are up 65.8% in nine months, and the European Union has agreed to cut Chinese hybrid exports by more than half. The inference is one step further: because the cheapest Chinese vehicles cannot legally enter the United States, their deflation is expressed in third-market prices and in a Mexican and Canadian pipeline rather than on American price lists. The speculation is the final step: if that pipeline matures, American and Japanese volume faces a floor that currently sits below a $27,904 compact car, and the companies best prepared for it are the ones that already licensed Chinese battery chemistry into US-built trucks. The collapse was designed into the subsidy formula — it just was not designed to travel.