Fiscal Policy & Stimulus

China's fiscal stance is still on hold — this batch again carries no central appropriation, no deficit revision and no special-purpose bond quota — but for the first time the demand data falsify the assumption that hold is sufficient. Retail sales rose just 1.1% y/y in January–September, to 370,001bn yuan (370.0trn), and 1.1% again in September alone ¹. Growth held at 4.6% in the first three quarters (Q1 5.0%, Q2 4.3%, Q3 4.5%, quarter-on-quarter +1.2%) ², but it is being carried by industry (+5.3%) and a strong export line, not by households. The gap between the announced posture and the delivered stimulus is now visible in the accounts.

The single most fiscal-relevant number in the batch is buried in the retail table. Headline sales rose 1.1% but sales excluding autos rose 2.8%, and in September ex-auto sales rose 3.9% against a 1.1% headline ¹. By subtraction from NBS's own published totals, auto retail fell roughly 21% y/y in September and 17–18% over January–September. Autos are precisely the category the trade-in and scrappage subsidy programme was built to move. This is the first hard evidence of what happens when that line lapses: the category reverses, and it drags the whole retail aggregate down with it. Nothing in this batch replaces the funding.

The household survey says the same thing from the income side. Per-capita disposable income rose 5.3% nominally and 4.4% in real terms in the first three quarters, but per-capita consumption expenditure rose only 3.6% nominally and 2.7% real — a gap of roughly 1.7 percentage points that is being saved, not spent ³. Property income, at +1.6%, was the weakest component; transfer income grew 5.7% and is already 18.6% of the total, which is the existing social-security baseline doing its work rather than a new stimulus. Consumption deflation removes any price-based urgency: September CPI was +0.9% with food −1.7%, pork −10.0% and services only +0.8% ⁴.

Composition confirms the model. Department stores fell 2.9%, brand专卖店 (brand specialty stores) 10.3% and specialty stores 2.2%, while convenience stores rose 5.2% and supermarkets 4.1% — households trading down within the same basket ¹. The one clean growth line is online services retail, +5.3% to 56,411bn yuan, which is exactly where the services-promotion strategy is pointed: Pan Gongsheng's 'investing in things and investing in people' line and the Jiangsu elderly-care build-out both aim to grow measured service output without a cash transfer ^5,6.

The counter-signal is that where money is genuinely committed, it still goes upstream. National R&D spending reached 3,928.82bn yuan in 2025, up 8.2% and 2.80% of GDP, with basic research up 13.3% to 283.46bn ⁷. Meanwhile the obligations keep being pushed down without matching money: Jiangsu's 15 elderly-care measures mandate 20 sqm of community care space per 100 households in new compounds, at least 4,000 lift retrofits over the 15th Five-Year Plan, county-level centres and full 15-minute service circles by 2030, with no subsidy line attached ⁶. Hunan's 'enterprise service year' adds a quasi-fiscal obligation of its own — clearing arrears owed to private firms, containing factor costs and regularising enforcement — which is cash out the door for local budgets, not stimulus into demand ⁸. The MOF's own September study session on '15th Five-Year Plan' fiscal work produced no figure at all ⁹.

So the hold now has a cost that is measurable, and the December decisions carry real weight. If Q4 retail stays near 1.1% with autos down a fifth, the cheapest fix is a renewed trade-in appropriation for 2027; the structural fix is 'investing in people' acquiring an actual budget line rather than remaining a phrase in communiqués ¹⁰. Watch four markers: whether a Q4 MOF/NDRC notice extends trade-in funding into 2027 or the Central Economic Work Conference announces it; whether the 2027 draft budget carries a matching transfer for the unfunded facility standards now cascading to provinces; whether October retail confirms the September auto collapse; and whether services inflation stays under 1%, which would keep the deflation case for household transfers alive.