Chinese refiners cannot export refined products freely. Export volumes are governed by quotas allocated administratively, in batches, during the year. A refiner with product and an open arbitrage still cannot ship without quota, and a refiner holding quota has an incentive to use it before it expires.
This inverts the usual causality. In an open market, product exports respond to the arbitrage: when the destination pays enough to cover freight, cargoes move. With quotas, exports respond to the allocation, and the arbitrage becomes a secondary consideration. A quota issued when the arb is closed can still produce exports, because unused quota is a wasted asset.
The market consequence is that Chinese product exports are partly a policy variable, and quota announcements are traded as news in their own right. A large allocation is bearish regional product cracks because it adds supply; a small one is supportive. That is a genuine and predictable transmission from an administrative decision to a price.
There is a second-order effect on crude that is easy to miss. Quota that permits product export supports refinery runs, because a refiner can process more crude when it has an outlet for the products. So a product export quota decision is also a crude demand decision, transmitted through run rates with a lag of weeks.
The system also interacts with the independent sector, which receives its own crude import quota allocations separately. A refiner can therefore be constrained on the input side, the output side, or both, and the binding constraint changes through the year. That is a different operating environment from the margin-driven one that the rest of the refining world lives in.