What moved
Beijing has opened a retroactive hunt for unpaid tax on offshore gains, reaching back in some cases to 2000, alongside new rules taxing offshore trust income at 20 per cent. The motive is fiscal. Budget revenue fell 1.7 per cent to Rmb21.6tn in 2025, and land sale revenue — the funding base of local government — collapsed from Rmb8.7tn in 2021 to Rmb4.15tn. Enforcement, meanwhile, pays: individual income tax rose 11.5 per cent last year against overall tax growth of 0.8 per cent.
Why it matters for pricing
The campaign against individuals is not the story for a corporate treasury. The arithmetic behind it is. An authority that has lost half its land sale income, and has found that enforcement grows revenue fourteen times faster than the economy does, will apply that lesson wherever the base is large and the argument is technical. Related-party financing is both.
China already has the tools and has never been reluctant to use them. Related-party interest is deductible only within a 2:1 debt-to-equity safe harbour for non-financial enterprises, 5:1 for financial — and above that only where a thin capitalisation special issue file argues the position from first principles. Outbound service fees and royalties have long attracted a benefit test. Guarantee fees and cash pool interest sit in exactly the same category: outbound, deductible, and priced against a benchmark almost nobody re-runs.
The second half of the announcement has had less attention and matters more. Beijing is tightening control of outbound capital flows at the same time. For a group with Chinese participants in a cross-border cash pool, or upstream lending out of China, that is a structural question rather than a compliance one. Cash that cannot leave is not pool liquidity. It is a local balance that still needs a local credit rate, and it no longer funds the treasury function the pricing assumed.
Do not expect the certainty procedures to absorb this. Across all jurisdictions, a transfer pricing MAP case took 30.9 months to close in 2024, and a bilateral APA 39.6 months to agree, up from 36.8. More telling: APAs rejected or closed without agreement rose from 12 per cent to over 19 per cent. Those are global averages, not Chinese figures. But they describe competent authorities under fiscal pressure conceding less — and China is the clearest case of that pressure anywhere.
Our call. Enforcement broadens from individuals to corporates within twelve to eighteen months, with interest deductibility and outbound guarantee fees the first levers. Assume audit before agreement, and assume a longer wait if you go to MAP.
What this means for your business
- Test your China related-party debt against the 2:1 ratio before someone else does. Above it, the interest is not deductible without a thin capitalisation file — and one written after the audit notice reads as a defence, not a policy.
- Re-run the benchmark on outbound guarantee fees and cash pool interest from China. These carry the weakest files in most groups, because they were priced once and inherited ever since.
- Model what happens to your pool if RMB balances cannot leave. Trapped cash is a local balance needing a local credit rate — not liquidity, and not what your pricing describes.
- If certainty matters, start now. Three years to a bilateral APA means an application filed in 2026 lands in 2029. One filed after the audit opens is not a plan.
Sources
Financial Times reporting on China’s offshore tax campaign and offshore trust rules, 5 August 2026, including Ministry of Finance and Wind revenue data. Thin capitalisation ratios and outbound payment scrutiny: PwC Worldwide Tax Summaries, People’s Republic of China. MAP and APA timings: OECD 2024 Mutual Agreement Procedure and Advance Pricing Arrangement Statistics, released October 2025. Figures as at 5 August 2026.