Which sectors will the 15th Five-Year Plan fund— and how much?

The cascade from doctrine to subsidy is legible.

The 14th Five-Year Plan, adopted by the NPC on 11 March 2021, named seven frontier fields, designated NEVs as a strategic emerging industry, and committed to building a national integrated big-data center system.¹ Across the five-year window that followed, those plan lines compiled into specific ministerial documents: Big Fund III (RMB 344bn, May 2024); the NEV purchase-tax extension to end-2027 (财政部公告 2023年第10号); the East-Data-West-Compute project (NDRC/CAC/MIIT/NEA joint, Feb 2022); and the world's first model-layer regulatory regime (CAC 生成式人工智能服务管理暂行办法, Aug 2023).² The cascade is densely observable. The reading is not.

MERICS writes careful, slow-cycle long-form on plan implementation.³Trivium China reads ministerial follow-ons daily for paying clients behind a paywall. Stanford's DigiChina hosts academic forums on individual provisions; MacroPolo writes retrospectives. None of them ship a tool that takes a single 5YP line and emits a structured cascade prediction — ranked beneficiaries, a subsidy band, a ministerial lag, and a falsification criterion — with the precedents that grounded each call shown alongside.

The gap exists because the work falls between two professional cultures. The people who can build the retrieval pipeline rarely read the 5YPs in the original language; the people who read the 5YPs rarely build retrieval pipelines. The thesis of this paper is that fourteen plans worth of realized history is enough training signal to make the fifteenth a structured prediction problem rather than a commentary problem. v0 documents the first half — ten realized 14FYP cascades — and commits the second half — six falsifiable predictions on the 15FYP — with resolution windows in 2028–2030.

Four moving parts. Each one shipped in v0.

The 14FYP text (gov.cn, 13 March 2021) plus the ten documented ministerial follow-ons indexed in RESEARCH.md §1.1–1.10. Each cascade carries a plan line, the realized ministerial document with its document number where one exists, the realized subsidy band, the lag in months, and the ranked realized beneficiary sectors. v0 scope is deliberately the 14FYP only; the 13FYP and earlier plans land in v1 with the historical re-ranking.

Each input excerpt is encoded with a TF-IDF vectorizer (sklearn, 1–2 ngrams, sublinear TF, English stopwords) fit over the historical-cascade excerpt corpus. Cosine similarity selects the top-K nearest precedents. A keyword-rule predictor — the v0 stand-in for the eventual LLM call — emits a SectorCascadePrediction with ranked beneficiary sectors, a subsidy band, a predicted ministerial lag, and a citation pack tying each call to either the subject excerpt or one of the retrieved precedents. The LLM step is the only swap between v0 and v1.

Three baselines run alongside on every backtest fold. Random: uniform sector draw. Mode: always the highest-empirical-weight sector in the 30-sector taxonomy. Keyword match: count taxonomy-keyword hits inside the excerpt, return the top-N by hit count. None are removable. If the retrieval pipeline is not lifting over the keyword baseline, the scoreboard says so before we do.

Recall at top-3 for the realized beneficiary sector set. MAE on lag prediction in months. Subsidy-band accuracy as a strict equality (five classes: sub_1bn / 1_5bn / 5_20bn / 20bn_plus / non_monetary). All three metrics evaluated leave-one-out across the 10 14FYP cascades. Calibration via reliability diagram is deferred to v1 once the live predictions begin resolving (late 2028).

What we will and will not publish.

Primary sources. Every cited 14FYP plan line resolves to gov.cn, the issuing ministry, the NPC archive, or a State Council bulletin. Press estimates appear with explicit attribution and an [unverified] tag where a primary fiscal document does not yet exist.

Falsification, not narration.Every live prediction in §III carries a criterion written so resolution does not require subjective adjudication. We score against the criterion, not against an analyst's memory of what we meant.

De-compilation is data.Section 1.7 (common prosperity) documents quiet retraction. A plan line that compiles and then de-compiles is a recoverable signal, not an error in the corpus — the prior should encode it as a distinct outcome class.

No fabricated magnitudes. Where a subsidy quantum is asserted by secondary press but not yet anchored to a primary fiscal document, the field carries the press attribution and the unverified tag. We do not interpolate.

Excluded scope. Defense and dual-use cascades are out of the public prediction set on editorial-guardrail grounds (Section 7 of the Forward Indicators plan). Quantum is documented as a counter-example, not predicted forward.