WP39: phase 0-2 scaffold — data audit, frozen design, PIT protocol, client, features, tests · 9c1ba37 · wp39_uqo

WP39: phase 0-2 scaffold — data audit, frozen design, PIT protocol, client, features, tests

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

Simon-Pierre Boucher committed 18 h ago (Nov 2, 2026) Browse files

76 changed files +6,968 −0

added .env.example +11 −0

++# WP39 UQO — environment-based authorization. Never commit a real .env.
++# FinData OS (mandatory primary data source). Key file is read if FINDATA_API_KEY is unset.
++FINDATA_BASE_URL=https://findata.spboucher.ai/api/v1
++FINDATA_API_KEY=
++FINDATA_API_KEY_FILE=~/findata-secrets/api-key-wp39.txt
++# Scratch cache for licensed raw extracts (outside the repository, never committed)
++WP39_CACHE=~/wp39-cache
++# SEC EDGAR fair-access User-Agent (public data, no secret) for ticker->CIK->SIC crosswalk
++SEC_USER_AGENT="UQO WP39 research (mailto:simon-pierre.boucher@uqo.ca)"
++# Set to 1 to run the live API contract tests
++WP39_LIVE=0

added .gitignore +35 −0

++# Raw & processed data — never committed (see data/raw/README.md for provenance)
++data/raw/*
++!data/raw/README.md
++data/processed/*
++!data/processed/.gitkeep
++
++# Python
++__pycache__/
++*.pyc
++.pytest_cache/
++.venv/
++*.egg-info/
++
++# LaTeX build artifacts (paper/main.pdf IS committed)
++paper/*.aux
++paper/*.log
++paper/*.out
++paper/*.toc
++paper/*.lof
++paper/*.lot
++paper/*.fls
++paper/*.fdb_latexmk
++paper/*.bbl
++paper/*.blg
++paper/*.spl
++paper/*.synctex.gz
++paper/*.run.xml
++paper/*.bcf
++
++# Logs
++logs/
++*.log
++
++# macOS
++.DS_Store

added CLAUDE.md +43 −0

++# CLAUDE.md — UQO Research WP39
++
++## THE INVISIBLE BALANCE SHEET
++### Can Intraday Financial Markets Detect Corporate Deterioration Before Accounting Disclosures?
++
++**Project ID:** `wp39_uqo` · **Institution:** Université du Québec en Outaouais (UQO) · **PI:** Simon-Pierre Boucher
++**Status:** research project built 2026-11-02 under the owner's specification (reproduced below in condensed form; the
++owner's original brief is the authoritative text). All findings are hypotheses until reproducibly estimated.
++
++Owner instructions added during the build (2026-11-02): (1) the FinData **fundamentals API** is the accounting source and
++must be used; (2) **minimise AI usage** — no LLM features, no neural/text models, no AI-generated statistics.
++
++## Non-negotiable execution rules
++1. FinData OS (`https://findata.spboucher.ai/api/v1`) is the primary source of prices, volumes, earnings events and
++   fundamentals. No silent substitution; every exception is documented in `DATA_AUDIT.md` / `docs/API_INTEGRATION.md`.
++2. Inspect before changing; preserve all other working papers; this repository is its own git repo.
++3. Credentials from the protected store only (`~/findata-secrets/api-key-wp39.txt` on the node); never printed, never
++   in URLs, git, artifacts, notebooks, figures or manuscript. No destructive ingestion, no unapproved backfills.
++4. FirstRate Data is privately licensed: no raw records in git, public endpoints, replication packages or logs.
++5. Hypotheses, variable definitions, PIT rules, baselines, splits and significance plan are frozen in
++   `docs/RESEARCH_DESIGN.md` and `docs/PIT_PROTOCOL.md` before interpreting results. Prediction ≠ causation.
++6. Every claim traces to a saved table/figure/log/test. No fabricated statistics or references.
++7. Phases with gates; `PROGRESS.md`, `DECISIONS.md`, `DATA_AUDIT.md`, `LIMITATIONS.md` are kept current.
++8. English everywhere; LaTeX article compiled with `latexmk` (`paper/`).
++
++## Where things are
++| Item | Location |
++|---|---|
++| Design, PIT rules, novelty | `docs/RESEARCH_DESIGN.md`, `docs/PIT_PROTOCOL.md`, `docs/NOVELTY_MEMO.md` |
++| Data audit, limitations, decisions, progress | root `*.md` files |
++| Code | `src/wp39/` (client, provenance, calendar, instruments, PIT fundamentals, corporate actions, ingest, features, labels, samples, splits, models, inference, portfolio, robustness, figures, pipeline) |
++| Pipeline | `scripts/00_…09_` in order; `scripts/01a_extract_node_tables.sh` runs on M3U96a |
++| Tests | `tests/` (`pytest`; live API tests need `WP39_LIVE=1`) |
++| Artifacts | `artifacts/manifests` (hashes, versions), `artifacts/evaluation`, `artifacts/logs`, `results/`, `figures/` |
++| Paper | `paper/main.tex` + `sections/` + generated `tables/`, `figures/` |
++
++## How to run (compute node M3U96a, API on localhost)
++```bash
++rsync -a --exclude .venv --exclude artifacts/logs ~/Desktop/UQO/Recherche/Recherche-UQO/finance/invisible-balance-sheet/ M3U96a:~/wp39_uqo/
++ssh M3U96a 'cd ~/wp39_uqo && uv venv -p 3.12 .venv && uv pip install -e ".[dev]" && export FINDATA_BASE_URL=http://127.0.0.1:8600/api/v1 && \
++  bash scripts/01a_extract_node_tables.sh && for s in 00 01 02 03 04 05 06 07 08; do .venv/bin/python scripts/${s}_*.py; done'
++```
++Then `rsync` `results/ figures/ artifacts/ paper/tables paper/figures` back and `cd paper && make`.

added DATA_AUDIT.md +167 −0

++# DATA_AUDIT.md — WP39 `wp39_uqo`
++
++Phase 0 discovery and feasibility audit of the **FinData OS** API (`https://findata.spboucher.ai/api/v1`) as the primary
++source for *The Invisible Balance Sheet*. All numbers below were measured on **2026-11-02** (America/Toronto) with the
++owner's Claude Code key (admin scope, laptop) and then with a dedicated least-privilege key `wp39_uqo`
++(scopes `read:market, read:fundamentals, query, export`; stored outside git in `~/findata-secrets/api-key-wp39.txt` on
++M3U96a and `~/Desktop/Cluster/secrets/findata/api-key-wp39.txt` on the laptop, mode 600). No credential appears in
++this repository, in logs or in artifacts. Probe transcripts: `artifacts/logs/phase0_probe.json` (produced by
++`scripts/00_probe_findata.py`).
++
++## 0. Viability conclusion
++
++**YES — viable, with two structural caveats that shape the paper.**
++
++1. A licensed, point-in-time usable cohort exists: **5,451 US-listed entities** with quarterly fundamentals
++   (**67 metrics**, `filing_date` 2012-01-03 → 2026-10-30, median 45 quarterly periods per firm), **373,326 earnings-release
++   events** with release date and (for 65 %) before-open/after-close timing, and **6.70 billion raw 1-minute bars** for
++   16,195 stock symbols (2000-01-03 → 2026-10-30) including 8,069 `-DELISTED` symbols. Exports of 1-minute history are
++   fast (≈ 1 s per symbol-decade, see §6).
++2. **Caveat A — survivorship in fundamentals, not in prices.** The fundamentals/earnings tables contain **zero delisted
++   entities** (ingestion ran for *active* tickers only). Minute bars do cover delisted names, so the price side is
++   survivorship-free, but the accounting panel is conditioned on being listed on 2026-11-01. The vendor does serve
++   fundamentals for at least some delisted tickers (probe: `FMD` → HTTP 200 with data; `CWZ` → 404), so a scoped
++   delisted-fundamentals backfill is *possible* but requires owner approval and a capacity plan (not run; see §8).
++3. **Caveat B — PIT is a proxy.** `available_at = filing_date + 1 day` is a policy, `revision_id = 1` for every fact
++   (`observed_at` = first ingestion, 2026-11-01/02). Values are the vendor's *current* history, so restatements that
++   happened before ingestion are invisible. The design therefore (i) uses only prior-quarter fundamentals as inputs,
++   (ii) uses the **earnings-release date** (not the filing date) as the reveal date, and (iii) reserves a stringent
++   specification using SEC-filing-level reconstruction for a subset (Phase 5, conditional on time).
++
++## 1. API contract — documented vs observed
++
++| Purpose | Route (observed in `openapi.json`, OpenAPI 3.1.0, API 1.0.0) | Observed behaviour |
++|---|---|---|
++| Datasets | `GET /catalog/datasets` | 64 datasets; 62 `active`, `fundamentals.sec_companyfacts` = `pending_configuration`, `market_ticks` = `unsupported` |
++| Dataset detail | `GET /catalog/datasets/{id}` | version, `license_class = frd-private` |
++| Bars | `GET /market/bars` | `interval ∈ {1m, 5m, 1d}` only (no 30m/1h on REST); `adjustment ∈ {raw, split, split_div}`; `limit ≤ 10 000`; `format ∈ {json, arrow, csv}`; `version` pin |
++| Bars schema | `GET /catalog/schema/bars` | schema v1, `timestamp_convention = bar_open`, `timezone_storage = UTC`, 20 fields (see §3) |
++| Fundamentals | `GET /fundamentals/{entity}` | `as_of`, `metrics`, `period_type ∈ {q,a}`, `include_superseded`, **`limit ≤ 20 000`** (default 2 000) |
++| Fundamental periods | `GET /fundamentals/{entity}/periods` | one row per fiscal period with `filing_date`, `available_at`, `max_revision`, `restated` |
++| SQL | `POST /query` | one read-only statement, `max_rows ≤ 100 000`, 30 s; views = dataset ids with `.`→`_` |
++| Exports | `POST /exports` → `GET /exports/{id}` → `/download` | body `{dataset_id, symbols ≤ 500, start, end, adjustment, format}`; **range ≤ 12 years per request (HTTP 413 otherwise)**; Parquet with schema metadata and pinned `X-Dataset-Version` |
++| Coverage | `GET /coverage`, `/coverage/symbols` (`limit ≤ 5 000`, cursor), `/coverage/heatmap` | per-session bar counts vs `expected_bars` (391 regular) |
++| Instruments | `GET /instruments`, `/instruments/{symbol}` | status `active`/`delisted`, first/last date |
++| Quality | `GET /quality`, `/quality/anomalies` | row-level flags (`price_jump`, `outside_session`, …) |
++| **Earnings events** | **no route** | dataset `fundamentals.earnings` is listed (0 rows reported) but rows live in Postgres `earnings_events` → read-only extraction on the node (documented exception, §7) |
++| Corporate actions | **no route** | Postgres `corporate_actions`: 6,131 splits (2000-01-03 → 2026-10-15), 243,209 cash dividends; applied at query time via `adjustment=` |
++
++Reconciliation notes: the catalog reports `row_count = 0` for Postgres-backed datasets (`fundamentals.*`, `corporate_actions`,
++`reference.instruments`); that is a catalog limitation, not an absence of data. The REST `interval` enum is narrower than the
++lake (30m/1h exist only as vendor-adjusted datasets); WP39 rebuilds 5m/30m/1d bars **from the raw 1m export** so that every
++frequency shares one source and one adjustment path.
++
++## 2. Market bars — coverage
++
++Counts of distinct stock symbols with at least one bar on a mid-June session (SQL playground, `session_date =` predicate):
++
++| Session | 1m symbols | 1m bars | 1d symbols | of which `-DELISTED` in 1d |
++|---|---:|---:|---:|---:|
++| 2001-06-15 | — | — | 1,088 | 0 |
++| 2005-06-15 | 3,498 (340,665 bars delisted) | 631,260 | 1,504 | 0 |
++| 2010-06-15 | — | — | 2,309 | 0 |
++| 2015-06-15 | 6,489 | 1,207,920 | 2,985 | 0 |
++| 2020-06-15 | — | — | 4,144 | 0 |
++| 2025-06-16 | 6,918 | 1,317,343 | 6,485 | 0 |
++
++Finding: the vendor **1d dataset is incomplete** relative to the 1m dataset (about half the symbols in 2005–2015 and no
++delisted names). Decision D-003: daily bars are *derived* from the 1m export (regular session 09:30–16:00 ET) — never taken
++from `market_bars.stock.1d.raw`. Delisted symbols are present in the 1m lake (e.g. `AA.B-DELISTED` 531 sessions
++2014–2016; `CWZ-DELISTED` 2000–2014).
++
++Bar conventions verified on `XOM` and `SPY` exports: timestamps are bar **open** in UTC (`2012-01-03 09:01Z` = 04:01 ET),
++extended hours 04:00–19:59 ET are present, zero-volume minutes are absent (bar counts per session 405–798 vs 391 expected
++regular bars), `quality_flags` list is empty on normal rows, `revision_id = 1`.
++
++## 3. Bars schema (Arrow, schema v1)
++
++`instrument_id, symbol_as_reported, venue (null), asset_class, interval, timestamp_utc (timestamp[us, UTC], bar open),
++session_date (date32, exchange tz), open, high, low, close, volume (double), currency, trade_count (int32?, derived only),
++vwap (double?, derived only), adjustment_type, source_id, ingested_at, quality_flags (list<string>), revision_id (int32)`.
++Parquet exports carry metadata `findata.schema = bars`, `findata.schema_version = 1`, `findata.timestamp_convention = bar_open`.
++
++## 4. Fundamentals — coverage and PIT fields
++
++- Entities: **5,451** (`frd:<TICKER>`, `security_id = firstrate:stock:<TICKER>`, none delisted, none with `security_id` null).
++- Metrics observed per period: **67** (the vendor documents 76; 9 are never populated in the API output). Names:
++  `acc_oci, assets_c, assets_nc, assets_tot, capex, cash, cash_business_acq, cash_change, cash_debt_sec, cash_div,
++  cash_equity, cash_fin, cash_inv, cash_inv_acq, cash_op, cogs, cos_inc, debt, debt_c, debt_nc, def_rev, depr, dps, ebit,
++  ebitda, ebt, eps, eps_d, equity, fx_impact_cash, gm, gp, intangibles, interest_exp, intr_cover, inv_c, inv_nc, inventory,
++  investments, liab_c, liab_nc, liab_tot, net_inc, net_inc_com, net_inc_non_control, net_loss_discontinued, om, op_ex,
++  op_inc, payables, ppe, pref_divs, receivables, research_dev, ret_earnings, rev, roa, roe, roic, sgna, share_comp, shares,
++  shares_av, shares_av_diluted, tax_asset, tax_expense, tax_liability`.
++- Quarterly `rev` facts by fiscal year (entities / facts): 2011: 1,798 / 1,812 · 2012: 2,461 / 8,725 · 2013: 2,623 / 9,868 ·
++  2014: 2,771 / 10,560 · 2015: 2,916 / 11,145 · 2016: 3,076 / 11,656 · 2017: 3,276 / 12,340 · 2018: 3,505 / 13,209 ·
++  2019: 3,763 / 14,034 · 2020: 4,309 / 15,488 · 2021: 4,663 / 17,458 · 2022: 4,851 / 18,289 · 2023: 5,043 / 18,904 ·
++  2024: 5,244 / 19,467 · 2025: 5,370 / 19,625 · 2026: 4,928 / 10,915. `filing_date` is never null (no `pit_fallback`).
++- Firms by number of quarterly periods: ≥ 8: 5,018 · ≥ 20: 4,331 · ≥ 40: 2,891. First quarter year: 2011: 2,141 ·
++  2012: 317 · 2013–2019: 148–256 per year · 2020: 551 · 2021: 353 · 2022–2025: 164–215 per year.
++- Filing lag (filing_date − fiscal_period_end, quarterly): p10 = 30 d, median = 40 d, p90 = 77 d.
++- Q4 values are genuine quarterly flows (AAPL FY2019 revenue 260.17 bn = 84.31 + 58.02 + 53.81 + 64.04 bn).
++- PIT fields per fact: `filing_date`, `published_at` (null), `available_at = filing_date + 1 d 00:00 ET`, `observed_at`
++  (= 2026-11-01 18:24 ET for the first ingestion), `revision_id = 1`, `restatement_flag = false`, `superseded_at = null`.
++  The API `as_of` semantics (`available_at ≤ as_of` and (`revision_id = 1` or `observed_at ≤ as_of`)) were exercised:
++  AAPL `as_of = 2020-01-01` returns 2,000 facts (limit) and reports `facts_hidden_by_as_of = 2211`.
++- Entity master: `cik`, `sector`, `industry` are **null for all 5,451 entities** → external SEC crosswalk needed (§7).
++
++## 5. Earnings-release events (reveal dates)
++
++Postgres `earnings_events` (`entity_id, event_date, eps_estimate, eps_reported, release_time, source_id, observed_at`):
++373,326 rows, 5,451 entities, 2000-01-01 → 2027-04-29, one row per (entity, date). `release_time`: `after_close` 124,832 ·
++`before_open` 120,586 · null 127,908. For 2012–2024 the null share is 35–36 %; `eps_estimate` is null for 24–27 % of
++events (vendor estimate; vintage unknown → exploratory only). Row counts for 2025–2026 are inflated by forward-dated
++placeholders (2026: 35,361 events), which are dropped at the sample end (2026-10-30).
++
++Matching quarterly filings to the first release event in `[period_end, filing_date + 10 d]` (213,954 filings 2012→):
++96.3 % matched; 1,493 (0.7 %) have `filing_date < event_date`; 16,868 (7.9 %) have a lag > 30 d (likely mismatches or
++late filers); median `filing_date − event_date` = **1 day**. Rule (PIT_PROTOCOL §3): reveal date `A_q` = matched release
++date; a firm-quarter is **excluded** when no release is matched, when `filing_date < event_date`, or when the lag exceeds
++30 days; the prediction cutoff always satisfies `c < min(A_q, F_q)` in trading days.
++
++## 6. Throughput and capacity (measured on M3U96a, API on localhost)
++
++| Operation | Measured |
++|---|---|
++| Export 1m raw, 1 symbol, 2012-01-01→2019-12-31 (`XOM`) | ready in 1.0 s, 859,354 rows, 13.6 MB Parquet, dataset version 17375 |
++| Export 1m raw, 1 symbol, 2020-01-01→2026-10-30 | ready in 1.0 s, 884,983 rows, 15.4 MB |
++| Full-range export (2012→2026, 14.8 years) | **HTTP 413** — limit 12 years per request → two requests per symbol |
++| SQL cross-section (`session_date =` on 1m) | ≈ 3 s |
++| Coverage per symbol (`/coverage/symbols`) | ms, cursor-paged |
++| Fundamentals, one entity, all quarterly facts | one call (≤ 20 000 rows) |
++
++Capacity model for the main sample (≈ 4,300 firms with ≥ 8 quarters and 1m coverage): 2 exports × ~1.5 s ≈ 3.6 h of
++export wall-time (sequential; 4 concurrent streams bring it under 1 h), ≈ 30 MB raw Parquet per firm (≈ 130 GB if kept).
++Decision D-004: raw extracts are processed **streaming** (export → daily feature rows at 1m/5m/30m/1d → delete raw file,
++keep SHA-256 + row count in the manifest), so steady-state scratch usage stays below 5 GB. The storage node M3U96a
++(28 cores, 96 GB RAM, 319 GB free on the internal SSD, hard floor 60 GB for FinData) hosts the computation; the feature
++panel (≈ 4,300 firms × ~3,700 sessions × 4 frequencies) is < 2 GB.
++
++## 7. External sources and exceptions
++
++| Need | Source | Status | Why not FinData |
++|---|---|---|---|
++| Reveal dates / release timing | FinData Postgres `earnings_events` (read-only `psql` on M3U96a) | available | not exposed by the API (gap logged for the FinData backlog) |
++| CIK, SIC industry | SEC `company_tickers.json` (10,423 rows) + `submissions/CIK##########.json` (`sic`, `sicDescription`, `exchanges`, `formerNames`) | verified reachable (fair-access UA) | entity master has no CIK/sector |
++| Daily factors (H5 only) | Kenneth French library, FF5 2×3 daily + momentum daily | verified (HTTP 200, updated 2026-10-26) | not a FinData dataset |
++| Analyst consensus | vendor `eps_estimate` in `earnings_events` (24–27 % missing, vintage unknown) | exploratory only | no point-in-time consensus history available |
++| TAQ / quotes / order flow | none | **not available** (vendor supplies bars only) | bar-based proxies are *never* called order flow, spreads or informed trading |
++
++## 8. Gaps, risks and recommendations to the owner
++
++1. **Delisted fundamentals backfill (recommended, not run).** A one-off `fundamentals` job for the 8,069 `-DELISTED`
++   tickers (vendor returns data for at least part of them) would remove the survivorship conditioning of the accounting
++   panel. Estimated cost: ≤ 8,069 vendor calls (1 req/s policy ≈ 2.3 h), a few hundred MB in Postgres. Requires: owner
++   approval, `findata ingest fundamentals --delisted` (does not exist yet; connector change on M3U96a `~/apps/findata`).
++2. **Earnings events API route** — expose `fundamentals.earnings` through `/api/v1/earnings/{entity}` so that WP39 (and
++   future projects) stop reading Postgres directly.
++3. **SEC EDGAR connector** (`fundamentals.sec_companyfacts`, pending a User-Agent setting) would give filing-level
++   as-originally-reported values for the stringent PIT specification.
++4. API bug found: `POST /admin/keys` returns HTTP 500 (`api_keys_user_id_fkey` violation) when the caller is an API key
++   whose `user_id` is not in `users`; the CLI `findata create-key` works. Logged for the FinData backlog, not fixed here.
++
++## 9. License and privacy constraints applied
++
++FirstRate Data: perpetual private use, research and derived statistics allowed with attribution, **no redistribution of raw
++records**. Consequently: no raw bars or fundamentals are committed; `data/raw` stays empty; the replication package ships
++code, manifests (hashes, versions, counts), derived aggregate tables and figures only; the manuscript attributes FirstRate
++Data via FinData OS. SEC and Kenneth French data are public.

added DECISIONS.md +14 −0

++# DECISIONS.md — dated design and engineering decisions (WP39)
++
++| ID | Date (ET) | Decision | Reason | Consequence |
++|---|---|---|---|---|
++| D-001 | 2026-11-02 13:00 | Compute on **M3U96a** (FinData host) with a dedicated least-privilege key `wp39_uqo`; repository lives in the UQO Research workspace (`Recherche-UQO/finance/invisible-balance-sheet`, dir name follows the series convention; project id `wp39_uqo`) and is rsync'd to `M3U96a:~/wp39_uqo` for runs | API on localhost, lake co-located, 28 cores / 96 GB; secrets never leave `~/findata-secrets` / `~/Desktop/Cluster/secrets` | results and manifests synced back; raw cache stays on the node |
++| D-002 | 2026-11-02 13:05 | **Earnings release dates** (`earnings_events.event_date`) define the reveal date `A_q`; the filing date `F_q` defines structured-fact availability (`available_at = F_q + 1`) | press releases precede 10-Q filings by a median of 1 day; cutoffs must precede the *first* public reveal | cutoff rule `c < min(A_q, F_q)`; release rows read from Postgres (no API route) — documented exception |
++| D-003 | 2026-11-02 13:10 | Daily bars are **rebuilt from the 1-minute raw export** (regular session), not taken from `market_bars.stock.1d.raw` | vendor 1d dataset covers about half the symbols before 2016 and no delisted names | one source, one adjustment path for all four frequencies |
++| D-004 | 2026-11-02 13:30 | Streaming feature extraction (export → features → delete raw Parquet; manifest keeps SHA-256) | disk budget on the FinData host (hard floor 60 GB free) | re-runs re-export from the pinned dataset version |
++| D-005 | 2026-11-02 14:00 | Universe = all FinData fundamentals entities (survivor-conditioned); delisted-fundamentals backfill **proposed, not run** | vendor ingestion covered active tickers only; backfill needs owner approval + capacity plan (rule 3 of the brief) | survivorship caveat stated in abstract, data section, limitations; price side is survivorship-free |
++| D-006 | 2026-11-02 14:05 | **No neural (B7) nor text/multimodal (B8) models; no LLM anywhere in the pipeline** | owner instruction received mid-session on 2026-11-02 ("limite le plus possible l'utilisation AI"); also keeps the frequency comparison about information sets rather than capacity | ladder is B0–B6; LightGBM provides the nonlinear benchmark with equal budgets |
++| D-007 | 2026-11-02 14:10 | Primary target = adverse **operating-income** surprise vs a seasonal random walk with drift, scaled by total assets, threshold −0.5 % | no point-in-time consensus exists in FinData; operating income is less affected by one-off items than net income; Foster (1977) model is the classic honest expectation | consensus-based miss (vendor `eps_estimate`) is exploratory only |
++| D-008 | 2026-11-02 14:15 | Frequency sets B3/B4/B5 use **identical feature families** (one column per family) computed from bars of 30 m / 5 m / 1 m | compare information sets, not dimensionality | a supplementary "all frequencies" model is run in Phase 4 |
++| D-009 | 2026-11-02 14:20 | Literature verified with **Crossref metadata** (no AI summarisation); references without a verifiable DOI are cited without one (Grossman–Stiglitz 1980, Holm 1979) | rule 7 of the brief (no fabricated references) | `artifacts/logs/crossref_batch*.json` are the evidence |
++| D-010 | 2026-11-02 14:30 | Phase gates that require "owner approval" are recorded as **self-approved under the autonomous mandate** with the memo attached; the owner can veto any of them before Phase 6 | the owner asked for the whole project to be built in one autonomous session and is not available mid-task | each gate entry in PROGRESS.md names the memo and the date |

added LIMITATIONS.md +31 −0

++# LIMITATIONS.md — living list (WP39)
++
++## Data
++- **Survivor-conditioned accounting panel.** Fundamentals and earnings-release rows exist only for the 5,451 tickers active
++  on 2026-11-01. Firms that delisted during 2012–2026 (including those that deteriorated and failed) are absent from the
++  firm-quarter sample, although their minute bars exist. Base rates of severe deterioration are therefore understated and
++  H3/H5 effects for the most vulnerable firms are not identified. A delisted-fundamentals backfill is proposed (DATA_AUDIT §8).
++- **PIT is a documented proxy.** `available_at = filing_date + 1 day` is a policy; all facts are `revision_id = 1` observed
++  on 2026-11-01/02, i.e. the vendor's current (possibly restated) history. Original-as-reported values are not available.
++  Inputs use prior quarters only, so restatements affect the *measurement of inputs and targets*, not the direction of time.
++- **OHLCV bars, not ticks.** No quotes, no trades, no order flow. Bar-based statistics (down-bar volume share, bar
++  autocorrelation) are descriptive proxies and are never interpreted as informed trading or spreads.
++- Vendor merges ticker changes under the latest ticker (fully since 2022); CIK/SIC come from the SEC's *current*
++  `company_tickers.json`, so historical ticker reuse before 2022 is only partially auditable (`formerNames` check).
++- `release_time` is unknown for ~35 % of release events; we never use release-day bars, so this affects only the
++  descriptive timing exhibit.
++- The vendor `eps_estimate` has an unknown vintage and is missing for a quarter of events → exploratory only.
++- 1-minute coverage is thinner before ~2005 and for illiquid names; the sample starts in 2012 and requires ≥ 100 regular
++  bars per session, which excludes the most illiquid firm-days (a documented selection on liquidity).
++
++## Design
++- The event-time cohort conditions on the *existence* of a release in the next 65 trading days; a hypothetical observer
++  knows the approximate schedule but not the exact date. The fixed-calendar design is reported as the robustness check.
++- Prediction, not causation: a positive H1 shows that bar-level information is statistically informative; it does not
++  establish who traded on private information.
++- No B7/B8 (neural/text) models by owner instruction; the frequency comparison is therefore within linear and
++  gradient-boosted families only.
++
++## To be completed as results arrive
++- (Phase 3) achieved sample sizes, base rates, coverage gaps by year/size.
++- (Phase 5) which placebo/stress tests failed or were inconclusive.

added PROGRESS.md +22 −0

++# PROGRESS.md — WP39 `wp39_uqo`
++
++## Phase 0 — Discovery and feasibility (2026-11-02) — GATE PASSED
++- Inspected the UQO Research workspace (`Recherche-UQO/`, 39 papers, harmonisation standard); no `wp39` existed; created
++  `finance/invisible-balance-sheet/` from the template; git initialised on a dedicated repository (no other paper touched).
++- Probed FinData: `/health` ok, 64 datasets, OpenAPI 3.1 (45 routes), bars schema v1, fundamentals (67 metrics, 5,451
++  entities, 2012→2026), earnings events (Postgres), coverage endpoints, export throughput (≈ 1 s per symbol-decade;
++  12-year cap per request). Created least-privilege key `wp39_uqo`.
++- Verified external public sources: SEC `company_tickers.json` + `submissions` (CIK/SIC), Kenneth French daily factors.
++- Verified 94 bibliography entries on Crossref; wrote `docs/NOVELTY_MEMO.md`, `literature/RELATED_WORK_MATRIX.csv`.
++- Outputs: `DATA_AUDIT.md` (viability = YES with survivorship + PIT caveats), `configs/*.yaml`, `docs/API_INTEGRATION.md`,
++  `docs/COMPUTE_BUDGET.md`, `docs/PIT_PROTOCOL.md`, `docs/RESEARCH_DESIGN.md` (frozen), `DECISIONS.md` D-001…D-010.
++- Owner instructions received during the session: (1) the fundamentals API must be used — it is the accounting source;
++  (2) minimise AI usage — no LLM features, no neural models (D-006).
++- Blockers: none. Next safe step: Phase 1 gate memo, then code + tests (Phase 2).
++
++## Phase 1 — Research design and novelty — GATE (self-approved under the autonomous mandate, D-010)
++- `docs/RESEARCH_DESIGN.md` frozen 2026-11-02 14:40 ET; hypotheses H1–H5 with null/alternative, primary endpoint (H1,
++  h = 20, log loss, binary adverse operating-income surprise), decision rules, splits, inference, power plan.
++- Owner review pending; nothing in Phases 2–5 depends on test-period data until the Phase 3 gate.
++
++## Phase 2 — Data acquisition and verification — IN PROGRESS

added configs/data_contract.yaml +40 −0

++# Data contract — verified against the live FinData OS API on 2026-11-02 (see DATA_AUDIT.md)
++findata:
++  base_url: https://findata.spboucher.ai/api/v1
++  openapi: /openapi.json
++  datasets:
++    bars_1m: market_bars.stock.1m.raw          # raw, bar-open timestamps in UTC, 04:00-20:00 ET, zero-volume minutes absent
++    bars_1m_etf: market_bars.etf.1m.raw        # SPY used as market proxy
++    bars_1d: market_bars.stock.1d.raw          # INCOMPLETE vs 1m (fewer symbols) -> daily bars are rebuilt from 1m
++    fundamentals: fundamentals.financials      # Postgres-backed; GET /fundamentals/{entity}
++    earnings: fundamentals.earnings            # Postgres table earnings_events (no API route yet) -> documented exception
++  limits:
++    bars_1m_days_per_call: 31
++    rows_per_page: 10000
++    export_years_per_request: 12
++    export_symbols_per_request: 500
++    query_timeout_s: 30
++    query_max_rows: 100000
++    rate_limit_per_min: 600
++    fundamentals_limit_max: 20000
++  bars_schema_version: 1
++  timestamp_convention: bar_open
++  timezone_storage: UTC
++  exchange_tz: America/New_York
++  regular_session: ["09:30", "16:00"]        # the 16:00 bar is the closing print
++  extended_session: ["04:00", "20:00"]
++  license_class: frd-private
++pit:
++  available_at_policy: "filing_date + 1 day 00:00 America/New_York (vendor proxy; fallback period_end + 91 d flagged pit_fallback)"
++  revision_semantics: "revision_id=1 observed 2026-11-01/02; no pre-ingestion restatement history (vendor limitation)"
++  reveal_date: "earnings_events.event_date (press release), release_time in {before_open, after_close, null}"
++universe:
++  entity_source: fundamentals entities (5,451 active tickers as of 2026-11-01; NO delisted entities -> survivorship caveat)
++  min_quarters: 8
++  first_filing_year: 2012
++  last_session: 2026-10-30
++external:
++  sec_company_tickers: https://www.sec.gov/files/company_tickers.json
++  sec_submissions: https://data.sec.gov/submissions/CIK{cik:010d}.json
++  ff5_daily: https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/ftp/F-F_Research_Data_5_Factors_2x3_daily_CSV.zip
++  mom_daily: https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/ftp/F-F_Momentum_Factor_daily_CSV.zip

added configs/models.yaml +27 −0

++# Equal tuning budget across information sets. Hyper-parameters are tuned ONLY on the validation period.
++baselines:
++  B0: {type: base_rate}
++  B1: {type: logistic, inputs: [fundamentals, controls]}
++  B2: {type: logistic, inputs: [fundamentals, controls, daily]}
++  B3: {type: logistic, inputs: [fundamentals, controls, daily, intraday_30m]}
++  B4: {type: logistic, inputs: [fundamentals, controls, daily, intraday_5m]}
++  B5: {type: logistic, inputs: [fundamentals, controls, daily, intraday_1m]}
++logistic:
++  penalty: l2
++  C_grid: [0.01, 0.03, 0.1, 0.3, 1.0, 3.0]
++  standardize: train_only
++  impute: train_median
++lightgbm:
++  information_sets: [B1, B2, B3, B4, B5]
++  budget_trials: 24          # identical random-search budget for every information set
++  seed: 20261102
++  search_space:
++    num_leaves: [7, 15, 31, 63]
++    learning_rate: [0.01, 0.03, 0.05]
++    min_child_samples: [50, 100, 200]
++    feature_fraction: [0.5, 0.7, 0.9]
++    bagging_fraction: [0.6, 0.8, 1.0]
++    lambda_l2: [0.0, 1.0, 10.0]
++  n_estimators_max: 2000
++  early_stopping_rounds: 100
++neural: {enabled: false, reason: "owner instruction 2026-11-02: minimise AI usage; B7/B8 not run"}

added configs/research.yaml +42 −0

++# Frozen research design parameters (see docs/RESEARCH_DESIGN.md). Changing a value requires a DECISIONS.md entry.
++seed: 20261102
++sample:
++  start: 2012-01-01
++  end: 2026-10-30
++  min_quarters_per_firm: 8
++  min_price_usd: 1.0                 # at cutoff, raw close
++  min_regular_bars_per_day: 100      # sessions with fewer 1m regular-session bars are "sparse"
++  min_valid_days_in_window: 15       # out of the 20-trading-day feature window
++cutoffs:
++  event_time_horizons_td: [40, 20, 10, 5, 1]   # trading days before the earnings release date A_q
++  primary_horizon_td: 20
++  fixed_calendar: month_end                     # last trading day of each month
++  embargo_rule: "cutoff c < min(A_q, F_q); bars with session_date <= c only; fundamentals with available_at <= c only"
++feature_window:
++  short_td: 20
++  long_td: 60
++frequencies: [1d, 30m, 5m, 1m]
++targets:
++  primary:
++    name: adverse_opinc_surprise
++    variable: op_inc
++    expectation: seasonal_random_walk_with_drift   # E[X_q] = X_{q-4} + mean(X_{q-1-k} - X_{q-5-k}, k=0..3), PIT
++    scale: assets_tot (latest available)
++    threshold: -0.005                               # surprise <= -0.5% of total assets
++  continuous_primary: scaled_opinc_surprise
++  secondary:
++    - {name: roa_yoy_decline, rule: "roa_q - roa_{q-4} <= -0.01"}
++    - {name: ocf_innovation_negative, rule: "(cash_op_q - cash_op_{q-4})/assets <= -0.01"}
++    - {name: liquidity_deterioration, rule: "current ratio falls >= 10% yoy OR interest coverage falls below 1 from >= 1"}
++    - {name: eps_miss, rule: "eps_reported < eps_estimate (vendor estimate, exploratory)"}
++splits:
++  train: [2012-01-01, 2018-12-31]
++  validation: [2019-01-01, 2020-12-31]
++  test: [2021-01-01, 2026-10-30]
++  walk_forward: {first_test_year: 2016, embargo_quarters: 1}
++  firm_holdout_fraction: 0.2
++inference:
++  cluster: [entity_id, calendar_quarter]
++  bootstrap_blocks: calendar_quarter
++  n_bootstrap: 1000
++  multiple_testing: holm

added data/README.md +17 −0

++# data/
++
++No licensed raw records are stored in this repository.
++
++| Path | Content | Committed? |
++|---|---|---|
++| `data/raw/` | nothing (raw FirstRate bars live only in the ignored scratch cache `$WP39_CACHE` on the compute node) | no |
++| `data/processed/` | derived firm-quarter panel, feature tables and labels (Parquet). Derived aggregates only; still **not** committed (size, license caution) | no |
++| `artifacts/manifests/` | hashed dataset manifests: dataset ids, `X-Dataset-Version`, extraction dates, row counts, SHA-256 of extracts, code SHA | yes |
++
++Provenance rules (see `DATA_AUDIT.md`, `docs/API_INTEGRATION.md`):
++- Market bars and fundamentals come from the owner's FinData OS API (`https://findata.spboucher.ai/api/v1`), a private
++  lakehouse of FirstRate Data (license: private use, no redistribution of raw records).
++- Earnings-release calendar rows are read from the FinData Postgres table `earnings_events` on the storage node because the
++  API does not expose that dataset yet (documented exception).
++- External public augmentations: SEC EDGAR `company_tickers.json` and `submissions` (CIK, SIC), Kenneth French data library
++  (daily factors). Both are public, versioned by download date in the manifests.

added data/processed/.gitkeep +0 −0

added data/raw/README.md +4 −0

++# data/raw — intentionally empty
++Raw FirstRate Data records (via FinData OS) are licensed for private use and are never stored in this repository.
++They live only in the ignored scratch cache on the compute node (`$WP39_CACHE`), are deleted after feature extraction,
++and are described by hashed manifests in `artifacts/manifests/`.

added docs/API_INTEGRATION.md +41 −0

++# API_INTEGRATION.md — how WP39 talks to FinData OS
++
++## Authorization
++- Header `Authorization: Bearer <key>`; the key is read from `FINDATA_API_KEY` or the file `FINDATA_API_KEY_FILE`
++  (default `~/findata-secrets/api-key-wp39.txt`, mode 600). The client never logs headers, never puts the key in a URL,
++  and redacts any string starting with `fd_` in exceptions (`src/wp39/api_client.py::_redact`).
++- Dedicated key `wp39_uqo` (scopes `read:market, read:fundamentals, query, export`; **no admin**), created 2026-11-02 with
++  the node CLI `findata create-key` because `POST /admin/keys` returned HTTP 500 for API-key callers (bug logged in
++  DATA_AUDIT §8).
++- Compute runs on **M3U96a** against `http://127.0.0.1:8600/api/v1` (same host as the lake; the public URL
++  `https://findata.spboucher.ai/api/v1` is used from the laptop for probes only).
++
++## Client (`src/wp39/api_client.py`)
++- `httpx.Client` with 60 s timeout, retry with bounded exponential backoff (5 attempts, 0.5 → 8 s, jitter) on 429/5xx and
++  network errors; idempotent GETs only are retried automatically; export creation is retried only when no export id was
++  received.
++- Every response's `X-Request-ID`, `X-Dataset-Id`, `X-Dataset-Version` are recorded in the provenance log
++  (`artifacts/logs/api_calls.jsonl`, one line per call, no payloads).
++- Cursor iteration for `/market/bars` and `/coverage/symbols`; export polling (`/exports/{id}`) every 1 s up to 10 min.
++- Arrow/Parquet schema validation against `GET /catalog/schema/bars` (schema v1) before any file is accepted.
++
++## Calls used by the pipeline
++| Step | Call | Scope | Volume |
++|---|---|---|---|
++| 00 probe | `/health`, `/catalog/datasets`, `/catalog/schema/bars`, `/openapi.json` | read | 4 |
++| 01 universe | `/coverage/symbols?dataset_id=market_bars.stock.1m.raw` (cursor, 5,000/page), `/instruments/{symbol}` | read | ≈ 4 + N |
++| 02 PIT panel | `/fundamentals/{entity}?period_type=q&limit=20000` (all facts, local as-of join), `/fundamentals/{entity}/periods` | read:fundamentals | 2 × N entities |
++| 03 features | `POST /exports` (1m raw, ≤ 12 y, 1 symbol) ×2 per symbol, `/exports/{id}`, `/download`; SPY from `market_bars.etf.1m.raw` | export | 2 × N + 2 |
++| checks | `POST /query` (cross-section counts, spot checks ≤ 100 k rows) | query | few |
++
++Exception (documented): earnings-release rows are extracted once on M3U96a with
++`psql -d findata -c "\copy (select ... from earnings_events) to '…'"` (read-only). The extract's SHA-256, row count and
++timestamp are in `artifacts/manifests/earnings_events.json`.
++
++## Idempotent partitioning and manifests
++- Exports are cached per `(symbol, chunk)` as `$WP39_CACHE/bars/<symbol>__<start>__<end>.parquet` and deleted after the
++  feature rows are written; `artifacts/manifests/bars_manifest.jsonl` keeps `symbol, start, end, dataset_id,
++  dataset_version, row_count, bytes, sha256, exported_at, request_id`.
++- Fundamentals extracts: `$WP39_CACHE/fundamentals/<entity>.parquet` with the same manifest fields.
++- Re-running a script skips completed partitions (manifest lookup) and never overwrites a partition with a different
++  dataset version without logging a `version_drift` warning.

added docs/COMPUTE_BUDGET.md +16 −0

++# COMPUTE_BUDGET.md
++
++| Resource | Plan | Measured / status |
++|---|---|---|
++| Compute node | M3U96a (Mac Studio M3 Ultra, 28 cores, 96 GB RAM) — FinData API on localhost | load 1.5–1.9 before start |
++| Scratch | `~/wp39-cache` on the internal SSD (319 GB free; FinData hard floor 60 GB) | streaming design keeps < 5 GB |
++| API exports | ≈ 2 × 4,300 symbol-chunks, ≈ 1 s each server-side, 4 concurrent | pilot measured 1.0 s / 860 k rows |
++| Fundamentals calls | 5,451 × 2 | ms each |
++| Feature computation | Polars per symbol (≈ 1.7 M rows → 4 frequency tables), 8 worker processes | to be measured in the pilot |
++| Models | logistic ridge (seconds) ; LightGBM 24 trials × 5 information sets × (1 chronological + 11 walk-forward folds) ≈ 1,440 fits, early stopping; ≈ 2–4 h on 28 cores | — |
++| Bootstrap | 1,000 stationary-bootstrap draws over 23 test quarters per contrast — seconds | — |
++| LaTeX | laptop (TeX Live, `latexmk`) | available |
++| AI usage | none in the pipeline (no LLM features, no neural B7/B8); Claude Code used only as the engineer | owner instruction 2026-11-02 |
++
++Pilot (Phase 2 gate): 300 firms stratified by first-fundamentals year and size, full history → validates throughput,
++feature sanity and label base rates before the full run.

added docs/NOVELTY_MEMO.md +53 −0

++# NOVELTY_MEMO.md — the three closest papers and why WP39 is not them
++
++Prepared 2026-11-02 from publisher metadata verified on Crossref (`literature/RELATED_WORK_MATRIX.csv`,
++`artifacts/logs/crossref_batch*.json`). Full texts were not machine-summarised; the points of distinction below are
++based on the papers' stated designs as reported in their abstracts and our own reading of their published versions.
++
++## Closest paper 1 — Grégoire & Martineau (2022), *How Is Earnings News Transmitted to Stock Prices?*, JAR 60(1)
++- **What it does.** Decomposes price discovery around earnings announcements using TAQ quotes *and* trades, showing that
++  most of the reaction to after-hours announcements occurs in quotes with little trading.
++- **Strongest objection to WP39 from this paper.** "Information arrives at the announcement; what you call anticipation is
++  the pre-announcement drift already documented, and bars cannot even see quotes."
++- **How WP39 differs.** WP39 does not study the announcement reaction at all: all features end strictly *before* the
++  release and the outcome is the accounting number itself, not the return. We make no claim about quotes versus trades
++  (our data are OHLCV bars, stated plainly) and we ask a forecasting question — whether the *resolution* of pre-release
++  bar data carries incremental information about fundamentals — which their event-window decomposition does not address.
++
++## Closest paper 2 — Chen, Cho, Dou & Lev (2022), *Predicting Future Earnings Changes Using Machine Learning and Detailed Financial Data*, JAR 60(2)
++- **What it does.** Shows that gradient-boosted trees on detailed XBRL financial-statement items predict the direction of
++  next-year earnings changes better than parsimonious accounting models, and that the predictions are priced slowly.
++- **Strongest objection.** "Fundamentals-based ML already predicts earnings changes; adding market data is Ou–Penman with
++  a new model, and daily returns are known to lead earnings (Beaver–Lambert–Morse 1980; Kothari–Sloan 1992)."
++- **How WP39 differs.** Their inputs are accounting items; ours hold the accounting information set fixed (B1) and
++  measure the *marginal* value of market data at four sampling frequencies with equal model capacity (B2–B5, B6). The
++  stock-prices-lead-earnings literature establishes the daily/annual lead; WP39 tests whether anything remains at the
++  minute scale once daily information is in, with a strict PIT availability clock and a release-date reveal clock.
++
++## Closest paper 3 — Zhang, Zhang, Cucuringu & Qian (2023/2024), *Volatility Forecasting with Machine Learning and Intraday Commonality*, JFEC
++- **What it does.** Pools intraday realized-volatility features across many stocks to forecast volatility with ML, showing
++  that commonality in intraday patterns improves volatility forecasts.
++- **Strongest objection.** "Intraday features forecast volatility, and earnings surprises are volatility events, so your
++  deterioration signal is a volatility forecast in disguise."
++- **How WP39 differs.** Our target is an accounting quantity with a *sign* (adverse surprise), not a variance; the design
++  includes the volatility families explicitly in every frequency set so that the test isolates information beyond
++  predicted volatility (placebo: a symmetric |surprise| target must show the pattern a volatility channel implies, a signed
++  target must not). We also report directional asymmetry (negative semivariance, down-bar volume) as pre-specified
++  mechanism variables, not as the headline.
++
++## Other adjacent work and the recycling check
++- UQO WP15 (*Faster, Wider, More Fragile?*) studies intraday *shock propagation* across stocks on the same vendor data;
++  it never links bars to accounting outcomes. WP39 shares the data vendor and the realized-measure toolkit only; no
++  result or figure is reused.
++- Pre-announcement informed trading (Christophe–Ferri–Angel 2004; Berkman et al. 2009; Kaniel et al. 2012;
++  Hendershott–Livdan–Schürhoff 2015) uses trader-type data we do not have; our contribution is orthogonal: whether
++  *anonymous* OHLCV at fine resolution is informative, and whether the information is attributable to the frequency.
++- Patton & Verardo (2012) and Bollerslev–Li–Xue (2018) model announcement-day betas and volume-volatility; they are
++  announcement-window studies.
++- Earnings-prediction ML (Easton et al. 2024; Cao et al. 2024) uses accounting, analyst and text inputs, not intraday bars.
++
++## Verdict
++No published paper, to our verified knowledge, measures the marginal forecasting value of 1-minute versus 5-, 30-minute
++and daily OHLCV for *forthcoming accounting deterioration* with a point-in-time information set and matched model
++capacity. The question is therefore open. If the empirical answer is a precise null, that is itself informative given
++the practitioner belief that high-frequency activity "sees" fundamentals early.

added docs/PIT_PROTOCOL.md +75 −0

++# PIT_PROTOCOL.md — point-in-time rules for WP39
++
++Frozen 2026-11-02 before any model was estimated. Any change is an amendment logged in `DECISIONS.md`.
++
++## 1. Clocks
++
++All instants are converted to **America/New_York** for session logic and kept in UTC in storage. Trading days follow the
++XNYS calendar (`exchange_calendars`). For firm `i` and fiscal quarter `q`:
++
++| Symbol | Meaning | Source |
++|---|---|---|
++| `E_q` | fiscal period end | `fundamentals_facts.fiscal_period_end` |
++| `A_q` | earnings-release date (press release) — the **reveal date** of quarter-`q` results | `earnings_events.event_date`, first event in `[E_q, F_q + 10 d]` |
++| `τ_q` | release timing: `before_open`, `after_close`, or `unknown` | `earnings_events.release_time` |
++| `F_q` | SEC filing date of the 10-Q/10-K | `fundamentals_facts.filing_date` |
++| `V_q` | vendor-proxy availability of the structured facts | `available_at = F_q + 1 d 00:00 ET` |
++| `c` | prediction cutoff (a trading day; features use information up to the **close** of `c`) | design |
++
++## 2. The information set at cutoff `c`
++
++`I(c)` contains exactly:
++
++1. **Market bars** with `session_date ≤ c` (regular session 09:30–16:00 ET for the main specification; the 16:00 bar is the
++   closing print; extended hours 04:00–09:29 and 16:01–19:59 enter only the "extended" variants, never mixed).
++2. **Fundamentals** with `available_at ≤ c 00:00 ET`, i.e. `F_j + 1 ≤ c`. The latest such quarter is `j*(c)`, and the
++   model also receives the *age* `c − A_{j*}` in trading days. Because `V_q > F_q ≥ A_q > c` by construction (§3), quarter-`q`
++   facts can never enter `I(c)`.
++3. **Earnings-release history** with `event_date < c` (prior realized releases, used for the historical base-rate feature B0).
++4. **Static crosswalk** (CIK, SIC) — time-invariant by assumption; `formerNames` are checked for ticker reuse (Phase 2 audit).
++5. **Nothing else.** No consensus estimates for quarter `q` (vintage unknown), no index-membership backfills, no
++   contemporaneous factors.
++
++## 3. Cutoff rules
++
++- **Event-time design** (primary, H1–H4): for horizon `h ∈ {40, 20, 10, 5, 1}` trading days, `c = A_q − h` trading days.
++  Admissible only if `c < min(A_q, F_q)` *and* `c ≥ A_{q−1} + 1` (the previous release has already happened, so the
++  cohort window never straddles two reveals). With `τ_q = before_open`, bars of `A_q` contain the reaction and are excluded
++  (`c ≤ A_q − 1` guarantees it); with `τ_q = after_close` or `unknown` the same rule applies — we never use `A_q` bars.
++- **Fixed-calendar design** (H1 robustness, H5): `c` = last trading day of each month; the target is the *next* release
++  after `c` (first `A_q > c`), restricted to `A_q − c ≤ 65` trading days. The identity of "the next quarter" is known ex
++  ante; its date is not needed by the forecaster.
++- **Firm-quarter exclusions** (frozen): no matched release; `F_q < A_q`; `F_q − A_q > 30` calendar days; `E_q` not within
++  60–120 days of `E_{q−1}` (irregular fiscal calendars); quarter `q − 4` or `q − 1` not available at `c` (expectation
++  undefined); raw close at `c` < 1 USD; fewer than 15 valid sessions in the 20-day feature window.
++
++## 4. Label construction (no price-based labels)
++
++Primary: `S_q = (X_q − E_c[X_q]) / TA_{j*}`, `X = op_inc` (quarterly operating income), `TA` = total assets of the latest
++available quarter, `E_c[X_q] = X_{q−4} + (1/4) Σ_{k=0..3} (X_{q−1−k} − X_{q−5−k})` using only quarters available at `c`
++(seasonal random walk with drift; Foster 1977). `D_q = 1[S_q ≤ −0.005]`. The secondary targets are in
++`configs/research.yaml`. `X_q` itself is taken from the **first ingested revision** of quarter `q` (the only one available),
++which is the vendor's current value — see LIMITATIONS §PIT.
++
++## 5. Leakage battery (each item is a unit test in `tests/`)
++
++| # | Risk | Guard | Test |
++|---|---|---|---|
++| L1 | quarter-`q` facts used as inputs | as-of join on `available_at ≤ c` | `test_pit_leakage.py::test_no_future_fundamentals` |
++| L2 | selection by fiscal end instead of availability | join key is `available_at`, never `fiscal_period_end` | `test_pit_leakage.py::test_join_key_is_available_at` |
++| L3 | post-release bars in the feature window | `session_date ≤ c < A_q` | `test_pit_leakage.py::test_no_bars_after_cutoff` |
++| L4 | global standardisation / imputation / tuning on test data | transformers fit on train rows only; tuning on validation period only | `test_train_only_transforms.py` |
++| L5 | DST / timezone errors | UTC → ET conversion, 09:30 and 16:00 bars identified on both DST regimes | `test_timestamp_dsts.py` |
++| L6 | ticker reuse / identifier drift | crosswalk with effective dates; SEC `formerNames` audit | `test_instrument_ids.py` |
++| L7 | splits/dividends creating mechanical returns | split-adjusted returns from `corporate_actions` (ex-date rule), raw bars immutable | `test_corporate_actions.py` |
++| L8 | release timing assumed intraday | release-day bars never used; timing enters only as a descriptive variable | `test_labels.py::test_release_day_excluded` |
++| L9 | overlapping target windows across splits | purge + 1-quarter embargo at every split boundary | `test_splits.py` |
++| L10 | survivorship by construction | universe = all fundamentals entities (documented survivor conditioning), delisted bars kept where present | `01_audit_universe.py` report |
++| L11 | vendor ingestion date mistaken for knowledge date | `observed_at` is never used as a knowledge date; `available_at` only | `test_pit_leakage.py::test_observed_at_not_used` |
++| L12 | randomised placebo | labels shuffled within calendar quarter → gain must vanish; cutoffs shifted −45 td → H4 gain must vanish | `08_run_robustness.py` |
++
++## 6. Equivalence check of the local as-of join with the API
++
++For 25 random entities and 10 random `as_of` instants, the set of (metric, period) pairs returned by
++`GET /fundamentals/{entity}?as_of=` must equal the set produced by the local filter `available_at ≤ as_of` on the full
++extract. Implemented in `tests/test_api_contract.py::test_asof_equivalence` (live test, `WP39_LIVE=1`).

added docs/REPLICATION.md +38 −0

++# REPLICATION.md — how to reproduce WP39
++
++## Requirements
++- Python ≥ 3.12 with `uv` (`uv venv -p 3.12 .venv && uv pip install -e ".[dev]"`).
++- Access to FinData OS (owner-issued key with scopes `read:market, read:fundamentals, query, export`) exposed as
++  `FINDATA_API_KEY` or a mode-600 file at `FINDATA_API_KEY_FILE`. Without access, Phases 2–5 cannot run; the saved
++  evaluation artifacts in `artifacts/evaluation/` and `results/` allow Phase 6 (tables/figures/paper) to be regenerated.
++- Read-only `psql` access to the FinData Postgres database on the storage node for `scripts/01a_extract_node_tables.sh`
++  (earnings events, entities, corporate actions — not yet exposed by the API).
++- Internet access to `sec.gov` (CIK/SIC) and the Kenneth French library (daily factors).
++- TeX Live with `latexmk` for the paper.
++
++## Order of execution (compute node, API on localhost)
++```bash
++export FINDATA_BASE_URL=http://127.0.0.1:8600/api/v1 WP39_CACHE=~/wp39-cache
++bash scripts/01a_extract_node_tables.sh          # node extracts + manifests
++python scripts/00_probe_findata.py               # contract reconciliation -> artifacts/logs/phase0_probe.json
++python scripts/01_audit_universe.py              # universe.parquet, results/universe_audit.json
++python scripts/02_build_pit_panel.py             # quarters.parquet (labels), design freeze hash
++python scripts/03_build_features.py --workers 8  # streaming exports -> session features -> rolling windows
++python scripts/04_build_labels.py                # sample_event.parquet, sample_calendar.parquet, cascades
++python scripts/05_train_baselines.py             # B0-B5 logistic, ridge, firm hold-out, walk-forward
++python scripts/06_train_advanced.py              # LightGBM equal-budget, horizon ladder, OOF signals
++python scripts/07_evaluate.py                    # metrics, paired contrasts, GW tests, Holm, power
++python scripts/08_run_robustness.py              # placebos, subgroups, panel regressions, H5
++python scripts/09_make_manuscript.py             # tables, figures, results_macros.tex
++cd paper && make                                 # main.pdf
++pytest -q                                        # 29 tests; WP39_LIVE=1 adds the live API contract tests
++```
++Runtime on a Mac Studio M3 Ultra (28 cores): see `PROGRESS.md` for measured wall-times of each step.
++
++## What is and is not in the package
++- In: code, configs, docs, tests, manifests (hashes/versions/counts), derived evaluation tables, figures, paper.
++- Out: raw bars and fundamentals (FirstRate Data license), the per-symbol feature cache (derived from licensed data;
++  regenerable), model binaries.
++- Determinism: all seeds come from `configs/research.yaml` (`seed: 20261102`); LightGBM uses `n_jobs=12` with
++  deterministic histogram construction at the default settings, so results are reproducible to numerical precision on
++  the same architecture.

added docs/RESEARCH_DESIGN.md +111 −0

++# RESEARCH_DESIGN.md — preregistered specification (internal)
++
++Frozen **2026-11-02 14:40 ET**, before the first test-period observation was opened. Hash of this file at freeze is
++recorded in `artifacts/manifests/design_freeze.json` by `scripts/02_build_pit_panel.py`. Later changes are amendments in
++`DECISIONS.md` (dated, with reason) and never alter frozen primary endpoints.
++
++## 1. Question
++
++Does one-minute OHLCV activity contain **incremental** information about a firm's *forthcoming* adverse accounting
++outcome, beyond (i) point-in-time fundamentals, (ii) daily OHLCV, (iii) coarser intraday aggregates of the same bars, and
++(iv) calendar/industry/size controls? Anticipation (prices move before the release) is not the claim; the claim is about
++the marginal forecasting value of **sampling frequency** holding the information set, calendar coverage, universe,
++cutoff, horizon and training budget fixed.
++
++## 2. Hypotheses — null, alternative, primary endpoint, decision rule
++
++| | Null `H0` | Alternative `H1` | Primary endpoint | Decision rule |
++|---|---|---|---|---|
++| **H1** incremental information | OOS log loss of B5 (fundamentals + daily + 1m) ≥ that of B2 (fundamentals + daily) | B5 < B2 | Δ log loss (B5 − B2) on the untouched test period, 2021-01 → 2026-10, horizon h = 20 | 95 % block-bootstrap CI (blocks = calendar quarters, 1,000 draws) excludes 0 **and** Giacomini–White conditional test p < 0.05 (two-sided) |
++| **H2** temporal resolution | Δ(B5 − B2) ≤ Δ(B4 − B2) and ≤ Δ(B3 − B2) | 1m gain strictly exceeds 5m and 30m gains | paired Δ log loss B5 − B4 and B5 − B3 | both CIs exclude 0 after Holm correction over the 3 frequency contrasts |
++| **H3** heterogeneity | the B5 − B2 gain is equal across terciles of size, illiquidity, leverage | gain larger in small, illiquid, levered firms | interaction coefficient `β_HF×tercile` in the panel regression (§5) | Holm-corrected p < 0.05 for the pre-specified ordering |
++| **H4** timing | the B5 − B2 gain does not increase as h shrinks (40→20→10→5→1) | monotone increase | slope of Δ log loss on −h across horizons | sign test on the five ordered horizons + placebo cutoffs (−45 td) showing no gain |
++| **H5** pricing | the LFDS does not predict the announcement-window abnormal return nor post-cutoff factor-adjusted returns | it does | (a) regression of CAR[c+1, A_q+1] on LFDS with controls; (b) decile long-short FF5+MOM alpha with costs | CI excludes 0; economic value reported with turnover and costs regardless of significance |
++
++Primary endpoint for the headline sentence: **H1 at h = 20 on the binary primary target** with log loss. Brier, AUC,
++PR-AUC, calibration slope/intercept, MAE/RMSE/OOS-R² (continuous target) are secondary metrics reported for every model.
++The headline sentence will read: *"At a 20-trading-day horizon, on a 2012–2026 panel of U.S. listed firms with a strictly
++point-in-time information set, one-minute OHLCV features change the out-of-sample log loss of forecasts of an adverse
++operating-income surprise by [Δ, 95 % CI] relative to a daily-market-and-accounting benchmark; the gain is
++[consistent/inconsistent] with [mechanisms] and has [measured/absent] economic value."*
++
++## 3. Sample
++
++Universe: all FinData fundamentals entities (5,451; survivor-conditioned, see LIMITATIONS) with ≥ 8 quarterly periods
++and 1-minute bars; firm-quarters 2012-Q1 → 2026-Q2 with reveal dates ≤ 2026-10-30 and admissible cutoffs (PIT_PROTOCOL
++§3). Sector = SEC SIC division (2-digit), size = log market cap at `c` (raw close × latest available `shares`).
++
++Expected sample (to be reported, not assumed): ~4,300 firms × ~40 quarters ≈ 150–170 k firm-quarters before exclusions.
++
++## 4. Variables
++
++**Targets** — `configs/research.yaml`. Primary binary `D_q = 1[S_q ≤ −0.5 % of assets]`, primary continuous `S_q`
++(winsorised at 1/99 % on the training period).
++
++**PIT fundamentals (B1)** — latest available quarter at `c`: `rev` growth (yoy), gross margin `gp/rev`, operating margin
++`op_inc/rev`, `roa`, `net_inc/assets_tot`, `cash_op/assets_tot`, accruals `(net_inc − cash_op)/assets_tot`,
++`cash/assets_tot`, leverage `debt/assets_tot`, interest coverage `intr_cover` (clipped), current ratio `assets_c/liab_c`,
++net issuance (Δ`shares` yoy), capex/assets, R&D/rev, `S_{j*}` (last surprise), last-four-quarter surprise mean, loss
++indicator, log assets, log market cap, information age `c − A_{j*}`, fiscal-quarter dummies, sector dummies, year-quarter
++calendar dummies (in-sample only; for prediction, calendar effects are replaced by the trailing cross-sectional base rate).
++
++**Daily OHLCV (B2)** — from 1m regular-session bars aggregated to daily: 20- and 60-day cumulative split-adjusted returns,
++daily return volatility (20d), Parkinson range volatility, skewness and kurtosis of daily returns (60d), daily volume
++relative to 60-day median (20d mean and last 5d), Amihud illiquidity (60d), share of zero-return days, overnight
++(close-to-open) vs intraday (open-to-close) return decomposition (20d sums), max drawdown (60d), beta and residual
++volatility against SPY (60d daily).
++
++**Intraday families (B3 = 30m, B4 = 5m, B5 = 1m)** — the *same* families computed on bars of the given length,
++each aggregated per session then averaged (or summed) over the 20-day window:
++realized variance (RV), bipower variation (BPV), relative jump `max(0, RV − BPV)/RV`, negative semivariance share
++`RS⁻/RV`, signed-jump variation `(RS⁺ − RS⁻)/RV`, realized skewness and kurtosis, maximum absolute bar-return share
++of RV (concentration), first-order autocorrelation of bar returns, variance ratio `RV_freq / RV_daily`, intraday drift
++`(close − open)/open`, time-of-day volume profile (share of volume in the first and last 30 minutes), volume dispersion
++across bars (Gini), down-bar volume share (volume on bars with negative return / total — *a bar statistic, not order
++flow*), volume surprise relative to the firm's 60-day time-of-day seasonal profile, fraction of bars present (activity),
++RV on market-down bars vs market-up bars (asymmetry vs SPY), realized beta to SPY at that frequency, vol-of-vol
++(dispersion of daily RV within the window). At 30m the within-day counts are small (13 bars), so skewness/kurtosis and
++autocorrelation are computed on the pooled window.
++
++Note on frequency comparison: B3/B4/B5 have the **same number of feature columns** (one per family); dimensionality does
++not grow with frequency. A supplementary experiment (Phase 4) concatenates all three to test whether 1m dominates when
++coarser aggregates are also present.
++
++## 5. Models and inference
++
++Baseline ladder B0–B6 (`configs/models.yaml`). Logistic ridge for B1–B5 with `C` tuned on the validation period only;
++LightGBM with an identical random-search budget (24 trials, early stopping on validation) for every information set.
++B7/B8 (temporal neural, text) are **not run** (owner instruction to minimise AI use; declared in DECISIONS D-006).
++
++Panel regression for interpretability and H3:
++`S_{i,q} = α_i + δ_{s(i),t(q)} + β·LFDS_{i,c} + γ'X_{i,c} + ε_{i,q}`, where `LFDS` is the B5 out-of-fold predicted
++probability (walk-forward, so it is never fitted on the same quarter), with firm fixed effects and sector×year-quarter
++effects, two-way clustered SEs (firm, calendar quarter). H3 interactions use ex-ante terciles formed at `c`.
++
++Comparisons: Diebold–Mariano / Giacomini–White on per-quarter mean loss differentials with HAC; stationary block
++bootstrap over calendar quarters for paired CIs; Holm within each hypothesis family; model confidence set over the ladder
++as a supplementary exhibit.
++
++## 6. Validation protocol
++
++- Chronological: train 2012-01 → 2018-12, validation 2019-01 → 2020-12, test 2021-01 → 2026-10 (never opened until Phase 3
++  gate). Purge: training observations whose `A_q` falls after the validation start minus one quarter are dropped
++  (embargo = 1 calendar quarter) and likewise at the test boundary.
++- Walk-forward: expanding window, annual refits, first test year 2016, 1-quarter embargo; this provides the out-of-fold
++  LFDS for the panel regressions and the H4/H5 exercises.
++- Firm hold-out: 20 % of firms (hash of entity id) never seen in training, evaluated separately from the time hold-out.
++- Train-only transforms: winsorisation bounds, medians for imputation, standardisation, sector-time demeaning, tercile
++  cut-points — all computed on the training rows of the current fold.
++
++## 7. Power
++
++With ≈ 40 k test firm-quarters, base rate ≈ 20–30 %, and quarter-level clustering (23 test quarters), the minimum
++detectable paired Δ log loss is roughly 0.002–0.004 (to be computed from the validation-period variance of the loss
++differential and reported in `artifacts/evaluation/power.json` before the test set is opened).
++
++## 8. What would make us abandon or pivot
++
++If, on the validation period, B2 already captures all of B5 (Δ within ±0.001 and CI covering 0), H1 is likely null; we
++proceed to the test anyway (a precise null is a result) and devote Phase 5 to characterising *where* intraday information
++does and does not help. If the earnings-release matching proves unreliable (> 10 % irregular lags after cleaning), the
++event-time design is demoted and the fixed-calendar design becomes primary (amendment required).

added literature/RELATED_WORK_MATRIX.csv +25 −0

++key,authors,year,journal,doi_or_url,data_source,frequency,sample,research_question,identification,findings,point_of_distinction_for_WP39
++gregoire2022,Grégoire & Martineau,2022,Journal of Accounting Research,10.1111/1475-679X.12394,TAQ quotes and trades,intraday (quotes/trades),US earnings announcements,How is earnings news incorporated into prices?,event-window decomposition of price discovery,most price discovery for after-hours announcements occurs through quote updates with little trading,"WP39 uses only pre-release bars, forecasts the accounting outcome (not the return), never claims quote/trade mechanisms"
++chen2022,Chen; Cho; Dou; Lev,2022,Journal of Accounting Research,10.1111/1475-679X.12429,XBRL detailed financial statement items,annual,US firms,Can ML on detailed financials predict earnings changes?,OOS prediction + return tests,boosted trees beat parsimonious models; slow pricing,"WP39 holds accounting inputs fixed and measures the marginal value of market data by sampling frequency"
++zhang2023volatility,Zhang; Zhang; Cucuringu; Qian,2024,Journal of Financial Econometrics,10.1093/jjfinec/nbad005,intraday equity returns,intraday,US stocks,Does pooling intraday commonality improve ML volatility forecasts?,OOS forecasting,commonality improves volatility forecasts,"WP39 targets a signed accounting quantity, includes volatility families as controls, uses a signed vs absolute placebo"
++beaver1980,Beaver; Lambert; Morse,1980,Journal of Accounting and Economics,10.1016/0165-4101(80)90013-0,CRSP/Compustat,annual,US,Do prices contain information about future earnings?,regression of earnings changes on returns,prices lead earnings,"establishes the daily/annual lead; WP39 asks what remains at minute resolution"
++kothari1992,Kothari & Sloan,1992,Journal of Accounting and Economics,10.1016/0165-4101(92)90016-U,CRSP/Compustat,annual,US,Information in prices about future earnings (ERC),leading returns,ERCs rise with leading returns,"same"
++collins1994,Collins; Kothari; Shanken; Sloan,1994,Journal of Accounting and Economics,10.1016/0165-4101(94)90024-8,CRSP/Compustat,annual,US,Why is the contemporaneous return-earnings association low?,lack of timeliness vs noise,prices anticipate earnings several years ahead,"same"
++foster1977,Foster,1977,The Accounting Review,10.2308/tar-4493420,quarterly earnings,quarterly,US,Time-series properties of quarterly earnings,forecast evaluation,seasonal random walk with drift forecasts well,"WP39's expectation model for the surprise"
++bernard1989,Bernard & Thomas,1989,Journal of Accounting Research,10.2307/2491062,CRSP/Compustat,daily,US,Post-earnings-announcement drift,event study,drift after SUE,"WP39 studies pre-release information, not post-announcement drift"
++ou1989,Ou & Penman,1989,Journal of Accounting and Economics,10.1016/0165-4101(89)90017-7,Compustat,annual,US,Can financial statement analysis predict earnings and returns?,logit on ratios,yes,"WP39 B1 baseline follows this tradition"
++sloan1996,Sloan,1996,The Accounting Review,10.2308/tar-9608042309,Compustat/CRSP,annual,US,Do prices reflect accruals information?,hedge portfolio,accrual anomaly,"accruals are a B1 control"
++christophe2004,Christophe; Ferri; Angel,2004,Journal of Finance,10.1111/j.1540-6261.2004.00681.x,Nasdaq short sales,daily,US,Do short sellers anticipate earnings news?,pre-announcement short selling,yes,"trader-type data unavailable to WP39; WP39 tests anonymous OHLCV"
++berkman2009,Berkman; Dimitrov; Jain; Koch; Tice,2009,Journal of Financial Economics,10.1016/j.jfineco.2008.04.009,CRSP/IBES,daily,US,Differences of opinion and returns around announcements,event study,prices fall around announcements for high-disagreement stocks,"announcement-window return study"
++kaniel2012,Kaniel; Liu; Saar; Titman,2012,Journal of Finance,10.1111/j.1540-6261.2012.01727.x,NYSE CAUD individual trades,daily,US,Do individual investors trade informatively before earnings?,pre-event trading,yes partly,"trader-type data"
++hendershott2015,Hendershott; Livdan; Schürhoff,2015,Journal of Financial Economics,10.1016/j.jfineco.2015.03.007,institutional order flow,daily,US,Are institutions informed about news?,order flow before news,yes,"order flow unavailable; WP39 never labels bar statistics as order flow"
++patton2012,Patton & Verardo,2012,Review of Financial Studies,10.1093/rfs/hhs073,TAQ,intraday,US,Does beta move with news?,announcement-day realized beta,beta rises on announcement days,"announcement-day study"
++bollerslev2018volume,Bollerslev; Li; Xue,2018,Review of Economic Studies,10.1093/restud/rdy003,TAQ,intraday,US,Volume-volatility relation around public news,structural model,announcement effects on volume-volatility elasticity,"announcement-window study"
++weller2018,Weller,2018,Review of Financial Studies,10.1093/rfs/hhx137,TAQ + earnings,daily/intraday,US,Does algorithmic trading reduce information acquisition?,AT and pre-announcement price informativeness,AT reduces pre-announcement price informativeness,"closest mechanism paper: WP39 reports pre-release information by era as a mechanism exhibit"
++lee2012,Lee,2012,Review of Financial Studies,10.1093/rfs/hhr084,TAQ,intraday,US,Jumps and information flow,jump tests around news,jumps cluster at news,"jump proxies are WP39 features"
++easton2024,Easton; Kapons; Monahan; Schütt; Weisbrod,2024,The Accounting Review,10.2308/tar-2021-0478,Compustat,annual,US,Forecast earnings with k-NN,OOS forecasting,kNN competitive with analysts,"accounting inputs only"
++cao2024,Cao; Jiang; Wang; Yang,2024,Journal of Financial Economics,10.1016/j.jfineco.2024.103910,analyst + ML,monthly,US,Man vs machine in stock analysis,OOS forecasts,AI complements analysts,"no intraday data"
++gu2020,Gu; Kelly; Xiu,2020,Review of Financial Studies,10.1093/rfs/hhaa009,CRSP/Compustat characteristics,monthly,US,ML for return prediction,OOS R2,trees and NNs best,"methodological reference for equal-budget comparisons"
++giacomini2006,Giacomini & White,2006,Econometrica,10.1111/j.1468-0262.2006.00718.x,—,—,—,Conditional predictive ability tests,—,—,"inference for H1/H2"
++liu2015,Liu; Patton; Sheppard,2015,Journal of Econometrics,10.1016/j.jeconom.2015.02.008,TAQ,intraday,multi-asset,Does anything beat 5-minute RV?,forecast comparison,5-min RV hard to beat,"frequency comparison for variance measurement; WP39 asks the analogous question for fundamentals"
++boucher2026wp15,Boucher,2026,UQO Working Paper 15,—,FirstRate Data 1-minute bars,1-minute,US 1992-2026,Intraday shock propagation,event studies,shocks propagate narrowly,"same vendor; different question; nothing reused"

added literature/references.bib +1021 −0

++% WP39 UQO bibliography. Journal metadata verified on Crossref 2026-11-02 (artifacts/logs/crossref_batch*.json).
++
++@article{gregoire2022,
++  author = {Grégoire, Vincent and Martineau, Charles},
++  title = {{How is Earnings News Transmitted to Stock Prices?}},
++  journal = {Journal of Accounting Research},
++  year = {2022},
++  volume = {60},
++  number = {1},
++  pages = {261--297},
++  doi = {10.1111/1475-679x.12394}
++}
++
++@article{zhang2023volatility,
++  author = {Zhang, Chao and Zhang, Yihuang and Cucuringu, Mihai and Qian, Zhongmin},
++  title = {{Volatility Forecasting with Machine Learning and Intraday Commonality}},
++  journal = {Journal of Financial Econometrics},
++  year = {2024},
++  volume = {22},
++  number = {2},
++  pages = {492--530},
++  doi = {10.1093/jjfinec/nbad005}
++}
++
++@article{ball1968,
++  author = {Ball, Ray and Brown, Philip},
++  title = {{An Empirical Evaluation of Accounting Income Numbers}},
++  journal = {Journal of Accounting Research},
++  year = {1968},
++  volume = {6},
++  number = {2},
++  pages = {159},
++  doi = {10.2307/2490232}
++}
++
++@article{bernard1989,
++  author = {Bernard, Victor L. and Thomas, Jacob K.},
++  title = {{Post-Earnings-Announcement Drift: Delayed Price Response or Risk Premium?}},
++  journal = {Journal of Accounting Research},
++  year = {1989},
++  volume = {27},
++  pages = {1},
++  doi = {10.2307/2491062}
++}
++
++@article{bernard1990,
++  author = {Bernard, Victor L. and Thomas, Jacob K.},
++  title = {{Evidence that stock prices do not fully reflect the implications of current earnings for future earnings}},
++  journal = {Journal of Accounting and Economics},
++  year = {1990},
++  volume = {13},
++  number = {4},
++  pages = {305--340},
++  doi = {10.1016/0165-4101(90)90008-r}
++}
++
++@article{sloan1996,
++  author = {Sloan, Richard G.},
++  title = {{Do Stock Prices Fully Reflect Information in Accruals and Cash Flows About Future Earnings?}},
++  journal = {The Accounting Review},
++  year = {1996},
++  volume = {71},
++  number = {3},
++  pages = {289--315},
++  doi = {10.2308/tar-9608042309}
++}
++
++@article{ou1989,
++  author = {Ou, Jane A. and Penman, Stephen H.},
++  title = {{Financial statement analysis and the prediction of stock returns}},
++  journal = {Journal of Accounting and Economics},
++  year = {1989},
++  volume = {11},
++  number = {4},
++  pages = {295--329},
++  doi = {10.1016/0165-4101(89)90017-7}
++}
++
++@article{kyle1985,
++  author = {Kyle, Albert S.},
++  title = {{Continuous Auctions and Insider Trading}},
++  journal = {Econometrica},
++  year = {1985},
++  volume = {53},
++  number = {6},
++  pages = {1315},
++  doi = {10.2307/1913210}
++}
++
++@article{easley1987,
++  author = {Easley, David and O'Hara, Maureen},
++  title = {{Price, trade size, and information in securities markets}},
++  journal = {Journal of Financial Economics},
++  year = {1987},
++  volume = {19},
++  number = {1},
++  pages = {69--90},
++  doi = {10.1016/0304-405x(87)90029-8}
++}
++
++@article{andersen2003,
++  author = {Andersen, Torben G. and Bollerslev, Tim and Diebold, Francis X. and Labys, Paul},
++  title = {{Modeling and Forecasting Realized Volatility}},
++  journal = {Econometrica},
++  year = {2003},
++  volume = {71},
++  number = {2},
++  pages = {579--625},
++  doi = {10.1111/1468-0262.00418}
++}
++
++@article{barndorff2004,
++  author = {Barndorff-Nielsen, O. E.},
++  title = {{Power and Bipower Variation with Stochastic Volatility and Jumps}},
++  journal = {Journal of Financial Econometrics},
++  year = {2004},
++  volume = {2},
++  number = {1},
++  pages = {1--37},
++  doi = {10.1093/jjfinec/nbh001}
++}
++
++@article{gu2020,
++  author = {Gu, Shihao and Kelly, Bryan and Xiu, Dacheng},
++  title = {{Empirical Asset Pricing via Machine Learning}},
++  journal = {The Review of Financial Studies},
++  year = {2020},
++  volume = {33},
++  number = {5},
++  pages = {2223--2273},
++  doi = {10.1093/rfs/hhaa009}
++}
++
++@article{giacomini2006,
++  author = {Giacomini, Raffaella and White, Halbert},
++  title = {{Tests of Conditional Predictive Ability}},
++  journal = {Econometrica},
++  year = {2006},
++  volume = {74},
++  number = {6},
++  pages = {1545--1578},
++  doi = {10.1111/j.1468-0262.2006.00718.x}
++}
++
++@article{diebold1995,
++  author = {Diebold, Francis X. and Mariano, Roberto S.},
++  title = {{Comparing Predictive Accuracy}},
++  journal = {Journal of Business \& Economic Statistics},
++  year = {1995},
++  volume = {13},
++  number = {3},
++  pages = {253--263},
++  doi = {10.1080/07350015.1995.10524599}
++}
++
++@article{hansen2005,
++  author = {Hansen, Peter Reinhard},
++  title = {{A Test for Superior Predictive Ability}},
++  journal = {Journal of Business \& Economic Statistics},
++  year = {2005},
++  volume = {23},
++  number = {4},
++  pages = {365--380},
++  doi = {10.1198/073500105000000063}
++}
++
++@article{fama2015,
++  author = {Fama, Eugene F. and French, Kenneth R.},
++  title = {{A five-factor asset pricing model}},
++  journal = {Journal of Financial Economics},
++  year = {2015},
++  volume = {116},
++  number = {1},
++  pages = {1--22},
++  doi = {10.1016/j.jfineco.2014.10.010}
++}
++
++@article{bollerslev2020good,
++  author = {Bollerslev, Tim and Li, Sophia Zhengzi and Zhao, Bingzhi},
++  title = {{Good Volatility, Bad Volatility, and the Cross Section of Stock Returns}},
++  journal = {Journal of Financial and Quantitative Analysis},
++  year = {2020},
++  volume = {55},
++  number = {3},
++  pages = {751--781},
++  doi = {10.1017/s0022109019000097}
++}
++
++@article{patton2015,
++  author = {Patton, Andrew J. and Sheppard, Kevin},
++  title = {{Good Volatility, Bad Volatility: Signed Jumps and The Persistence of Volatility}},
++  journal = {Review of Economics and Statistics},
++  year = {2015},
++  volume = {97},
++  number = {3},
++  pages = {683--697},
++  doi = {10.1162/rest_a_00503}
++}
++
++@article{chen2022,
++  author = {CHEN, XI and CHO, YANG HA (TONY) and DOU, YIWEI and LEV, BARUCH},
++  title = {{Predicting Future Earnings Changes Using Machine Learning and Detailed Financial Data}},
++  journal = {Journal of Accounting Research},
++  year = {2022},
++  volume = {60},
++  number = {2},
++  pages = {467--515},
++  doi = {10.1111/1475-679x.12429}
++}
++
++@article{piotroski2000,
++  author = {Piotroski, Joseph D.},
++  title = {{Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers}},
++  journal = {Journal of Accounting Research},
++  year = {2000},
++  volume = {38},
++  pages = {1},
++  doi = {10.2307/2672906}
++}
++
++@article{hirshleifer2009,
++  author = {HIRSHLEIFER, DAVID and LIM, SONYA SEONGYEON and TEOH, SIEW HONG},
++  title = {{Driven to Distraction: Extraneous Events and Underreaction to Earnings News}},
++  journal = {The Journal of Finance},
++  year = {2009},
++  volume = {64},
++  number = {5},
++  pages = {2289--2325},
++  doi = {10.1111/j.1540-6261.2009.01501.x}
++}
++
++@article{dellavigna2009,
++  author = {DELLAVIGNA, STEFANO and POLLET, JOSHUA M.},
++  title = {{Investor Inattention and Friday Earnings Announcements}},
++  journal = {The Journal of Finance},
++  year = {2009},
++  volume = {64},
++  number = {2},
++  pages = {709--749},
++  doi = {10.1111/j.1540-6261.2009.01447.x}
++}
++
++@article{loughran2011,
++  author = {LOUGHRAN, TIM and MCDONALD, BILL},
++  title = {{When Is a Liability Not a Liability? Textual Analysis, Dictionaries, and 10-Ks}},
++  journal = {The Journal of Finance},
++  year = {2011},
++  volume = {66},
++  number = {1},
++  pages = {35--65},
++  doi = {10.1111/j.1540-6261.2010.01625.x}
++}
++
++@article{bajgrowicz2016,
++  author = {Bajgrowicz, Pierre and Scaillet, Olivier and Treccani, Adrien},
++  title = {{Jumps in High-Frequency Data: Spurious Detections, Dynamics, and News}},
++  journal = {Management Science},
++  year = {2016},
++  volume = {62},
++  number = {8},
++  pages = {2198--2217},
++  doi = {10.1287/mnsc.2015.2234}
++}
++
++@article{lee2008,
++  author = {Lee, Suzanne S. and Mykland, Per A.},
++  title = {{Jumps in Financial Markets: A New Nonparametric Test and Jump Dynamics}},
++  journal = {Review of Financial Studies},
++  year = {2008},
++  volume = {21},
++  number = {6},
++  pages = {2535--2563},
++  doi = {10.1093/rfs/hhm056}
++}
++
++@article{bollerslev2018risk,
++  author = {Bollerslev, Tim and Hood, Benjamin and Huss, John and Pedersen, Lasse Heje},
++  title = {{Risk Everywhere: Modeling and Managing Volatility}},
++  journal = {The Review of Financial Studies},
++  year = {2018},
++  volume = {31},
++  number = {7},
++  pages = {2729--2773},
++  doi = {10.1093/rfs/hhy041}
++}
++
++@article{hendershott2011,
++  author = {HENDERSHOTT, TERRENCE and JONES, CHARLES M. and MENKVELD, ALBERT J.},
++  title = {{Does Algorithmic Trading Improve Liquidity?}},
++  journal = {The Journal of Finance},
++  year = {2011},
++  volume = {66},
++  number = {1},
++  pages = {1--33},
++  doi = {10.1111/j.1540-6261.2010.01624.x}
++}
++
++@article{richardson2005,
++  author = {Richardson, Scott A. and Sloan, Richard G. and Soliman, Mark T. and Tuna, İrem},
++  title = {{Accrual reliability, earnings persistence and stock prices}},
++  journal = {Journal of Accounting and Economics},
++  year = {2005},
++  volume = {39},
++  number = {3},
++  pages = {437--485},
++  doi = {10.1016/j.jacceco.2005.04.005}
++}
++
++@article{beneish1999,
++  author = {Beneish, Messod D.},
++  title = {{The Detection of Earnings Manipulation}},
++  journal = {Financial Analysts Journal},
++  year = {1999},
++  volume = {55},
++  number = {5},
++  pages = {24--36},
++  doi = {10.2469/faj.v55.n5.2296}
++}
++
++@article{campbell2008,
++  author = {CAMPBELL, JOHN Y. and HILSCHER, JENS and SZILAGYI, JAN},
++  title = {{In Search of Distress Risk}},
++  journal = {The Journal of Finance},
++  year = {2008},
++  volume = {63},
++  number = {6},
++  pages = {2899--2939},
++  doi = {10.1111/j.1540-6261.2008.01416.x}
++}
++
++@article{fama1973,
++  author = {Fama, Eugene F. and MacBeth, James D.},
++  title = {{Risk, Return, and Equilibrium: Empirical Tests}},
++  journal = {Journal of Political Economy},
++  year = {1973},
++  volume = {81},
++  number = {3},
++  pages = {607--636},
++  doi = {10.1086/260061}
++}
++
++@article{petersen2009,
++  author = {Petersen, Mitchell A.},
++  title = {{Estimating Standard Errors in Finance Panel Data Sets: Comparing Approaches}},
++  journal = {Review of Financial Studies},
++  year = {2009},
++  volume = {22},
++  number = {1},
++  pages = {435--480},
++  doi = {10.1093/rfs/hhn053}
++}
++
++@article{cameron2011,
++  author = {Cameron, A. Colin and Gelbach, Jonah B. and Miller, Douglas L.},
++  title = {{Robust Inference With Multiway Clustering}},
++  journal = {Journal of Business \& Economic Statistics},
++  year = {2011},
++  volume = {29},
++  number = {2},
++  pages = {238--249},
++  doi = {10.1198/jbes.2010.07136}
++}
++
++@article{novymarx2016,
++  author = {Novy-Marx, Robert and Velikov, Mihail},
++  title = {{A Taxonomy of Anomalies and Their Trading Costs}},
++  journal = {Review of Financial Studies},
++  year = {2016},
++  volume = {29},
++  number = {1},
++  pages = {104--147},
++  doi = {10.1093/rfs/hhv063}
++}
++
++@article{bollerslev2016beta,
++  author = {Bollerslev, Tim and Li, Sophia Zhengzi and Todorov, Viktor},
++  title = {{Roughing up beta: Continuous versus discontinuous betas and the cross section of expected stock returns}},
++  journal = {Journal of Financial Economics},
++  year = {2016},
++  volume = {120},
++  number = {3},
++  pages = {464--490},
++  doi = {10.1016/j.jfineco.2016.02.001}
++}
++
++@article{lo1988,
++  author = {Lo, Andrew W. and MacKinlay, A. Craig},
++  title = {{Stock Market Prices Do Not Follow Random Walks: Evidence from a Simple Specification Test}},
++  journal = {Review of Financial Studies},
++  year = {1988},
++  volume = {1},
++  number = {1},
++  pages = {41--66},
++  doi = {10.1093/rfs/1.1.41}
++}
++
++@article{freyberger2020,
++  author = {Freyberger, Joachim and Neuhierl, Andreas and Weber, Michael},
++  title = {{Dissecting Characteristics Nonparametrically}},
++  journal = {The Review of Financial Studies},
++  year = {2020},
++  volume = {33},
++  number = {5},
++  pages = {2326--2377},
++  doi = {10.1093/rfs/hhz123}
++}
++
++@article{bollerslev2016errors,
++  author = {Bollerslev, Tim and Patton, Andrew J. and Quaedvlieg, Rogier},
++  title = {{Exploiting the errors: A simple approach for improved volatility forecasting}},
++  journal = {Journal of Econometrics},
++  year = {2016},
++  volume = {192},
++  number = {1},
++  pages = {1--18},
++  doi = {10.1016/j.jeconom.2015.10.007}
++}
++
++@article{liu2015,
++  author = {Liu, Lily Y. and Patton, Andrew J. and Sheppard, Kevin},
++  title = {{Does anything beat 5-minute RV? A comparison of realized measures across multiple asset classes}},
++  journal = {Journal of Econometrics},
++  year = {2015},
++  volume = {187},
++  number = {1},
++  pages = {293--311},
++  doi = {10.1016/j.jeconom.2015.02.008}
++}
++
++@article{andersen1998,
++  author = {Andersen, Torben G. and Bollerslev, Tim},
++  title = {{Answering the Skeptics: Yes, Standard Volatility Models do Provide Accurate Forecasts}},
++  journal = {International Economic Review},
++  year = {1998},
++  volume = {39},
++  number = {4},
++  pages = {885},
++  doi = {10.2307/2527343}
++}
++
++@article{aboody2010,
++  author = {Aboody, David and Lehavy, Reuven and Trueman, Brett},
++  title = {{Limited attention and the earnings announcement returns of past stock market winners}},
++  journal = {Review of Accounting Studies},
++  year = {2010},
++  volume = {15},
++  number = {2},
++  pages = {317--344},
++  doi = {10.1007/s11142-009-9104-9}
++}
++
++@article{kothari2001,
++  author = {Kothari, S.P},
++  title = {{Capital markets research in accounting}},
++  journal = {Journal of Accounting and Economics},
++  year = {2001},
++  volume = {31},
++  number = {1-3},
++  pages = {105--231},
++  doi = {10.1016/s0165-4101(01)00030-1}
++}
++
++@article{easley2012,
++  author = {Easley, David and López de Prado, Marcos M. and O'Hara, Maureen},
++  title = {{Flow Toxicity and Liquidity in a High-frequency World}},
++  journal = {Review of Financial Studies},
++  year = {2012},
++  volume = {25},
++  number = {5},
++  pages = {1457--1493},
++  doi = {10.1093/rfs/hhs053}
++}
++
++@article{hou2015,
++  author = {Hou, Kewei and Xue, Chen and Zhang, Lu},
++  title = {{Digesting Anomalies: An Investment Approach}},
++  journal = {Review of Financial Studies},
++  year = {2015},
++  volume = {28},
++  number = {3},
++  pages = {650--705},
++  doi = {10.1093/rfs/hhu068}
++}
++
++@article{dechow2010,
++  author = {Dechow, Patricia and Ge, Weili and Schrand, Catherine},
++  title = {{Understanding earnings quality: A review of the proxies, their determinants and their consequences}},
++  journal = {Journal of Accounting and Economics},
++  year = {2010},
++  volume = {50},
++  number = {2-3},
++  pages = {344--401},
++  doi = {10.1016/j.jacceco.2010.09.001}
++}
++
++@article{fama1970,
++  author = {Fama, Eugene F.},
++  title = {{Efficient Capital Markets: A Review of Theory and Empirical Work}},
++  journal = {The Journal of Finance},
++  year = {1970},
++  volume = {25},
++  number = {2},
++  pages = {383},
++  doi = {10.2307/2325486}
++}
++
++@article{bollerslev2011tails,
++  author = {BOLLERSLEV, TIM and TODOROV, VIKTOR},
++  title = {{Tails, Fears, and Risk Premia}},
++  journal = {The Journal of Finance},
++  year = {2011},
++  volume = {66},
++  number = {6},
++  pages = {2165--2211},
++  doi = {10.1111/j.1540-6261.2011.01695.x}
++}
++
++@article{hirshleifer2003,
++  author = {Hirshleifer, David and Teoh, Siew Hong},
++  title = {{Limited attention, information disclosure, and financial reporting}},
++  journal = {Journal of Accounting and Economics},
++  year = {2003},
++  volume = {36},
++  number = {1-3},
++  pages = {337--386},
++  doi = {10.1016/j.jacceco.2003.10.002}
++}
++
++@article{savor2016,
++  author = {SAVOR, PAVEL and WILSON, MUNGO},
++  title = {{Earnings Announcements and Systematic Risk}},
++  journal = {The Journal of Finance},
++  year = {2016},
++  volume = {71},
++  number = {1},
++  pages = {83--138},
++  doi = {10.1111/jofi.12361}
++}
++
++@article{budish2015,
++  author = {Budish, Eric and Cramton, Peter and Shim, John},
++  title = {{The High-Frequency Trading Arms Race: Frequent Batch Auctions as a Market Design Response *}},
++  journal = {The Quarterly Journal of Economics},
++  year = {2015},
++  volume = {130},
++  number = {4},
++  pages = {1547--1621},
++  doi = {10.1093/qje/qjv027}
++}
++
++@article{chen2024deep,
++  author = {Chen, Luyang and Pelger, Markus and Zhu, Jason},
++  title = {{Deep Learning in Asset Pricing}},
++  journal = {Management Science},
++  year = {2024},
++  volume = {70},
++  number = {2},
++  pages = {714--750},
++  doi = {10.1287/mnsc.2023.4695}
++}
++
++@article{kozak2020,
++  author = {Kozak, Serhiy and Nagel, Stefan and Santosh, Shrihari},
++  title = {{Shrinking the cross-section}},
++  journal = {Journal of Financial Economics},
++  year = {2020},
++  volume = {135},
++  number = {2},
++  pages = {271--292},
++  doi = {10.1016/j.jfineco.2019.06.008}
++}
++
++@article{brogaard2019,
++  author = {BROGAARD, JONATHAN and HENDERSHOTT, TERRENCE and RIORDAN, RYAN},
++  title = {{Price Discovery without Trading: Evidence from Limit Orders}},
++  journal = {The Journal of Finance},
++  year = {2019},
++  volume = {74},
++  number = {4},
++  pages = {1621--1658},
++  doi = {10.1111/jofi.12769}
++}
++
++@article{ang2006,
++  author = {ANG, ANDREW and HODRICK, ROBERT J. and XING, YUHANG and ZHANG, XIAOYAN},
++  title = {{The Cross-Section of Volatility and Expected Returns}},
++  journal = {The Journal of Finance},
++  year = {2006},
++  volume = {61},
++  number = {1},
++  pages = {259--299},
++  doi = {10.1111/j.1540-6261.2006.00836.x}
++}
++
++@article{skinner2002,
++  author = {Skinner, Douglas J. and Sloan, Richard G.},
++  title = {{Earnings Surprises, Growth Expectations, and Stock Returns or Don't Let an Earnings Torpedo Sink Your Portfolio}},
++  journal = {Review of Accounting Studies},
++  year = {2002},
++  volume = {7},
++  number = {2-3},
++  pages = {289--312},
++  doi = {10.1023/a:1020294523516}
++}
++
++@article{levi2015,
++  author = {Levi, Shai and Zhang, Xiao-Jun},
++  title = {{Asymmetric decrease in liquidity trading before earnings announcements and the announcement return premium}},
++  journal = {Journal of Financial Economics},
++  year = {2015},
++  volume = {118},
++  number = {2},
++  pages = {383--398},
++  doi = {10.1016/j.jfineco.2015.08.003}
++}
++
++@article{campbell1993,
++  author = {Campbell, J. Y. and Grossman, S. J. and Wang, J.},
++  title = {{Trading Volume and Serial Correlation in Stock Returns}},
++  journal = {The Quarterly Journal of Economics},
++  year = {1993},
++  volume = {108},
++  number = {4},
++  pages = {905--939},
++  doi = {10.2307/2118454}
++}
++
++@article{fang2009,
++  author = {FANG, LILY and PERESS, JOEL},
++  title = {{Media Coverage and the Cross-section of Stock Returns}},
++  journal = {The Journal of Finance},
++  year = {2009},
++  volume = {64},
++  number = {5},
++  pages = {2023--2052},
++  doi = {10.1111/j.1540-6261.2009.01493.x}
++}
++
++@article{beaver1980,
++  author = {Beaver, William and Lambert, Richard and Morse, Dale},
++  title = {{The information content of security prices}},
++  journal = {Journal of Accounting and Economics},
++  year = {1980},
++  volume = {2},
++  number = {1},
++  pages = {3--28},
++  doi = {10.1016/0165-4101(80)90013-0}
++}
++
++@article{kothari1992,
++  author = {Kothari, S.P and Sloan, Richard G},
++  title = {{Information in prices about future earnings}},
++  journal = {Journal of Accounting and Economics},
++  year = {1992},
++  volume = {15},
++  number = {2-3},
++  pages = {143--171},
++  doi = {10.1016/0165-4101(92)90016-u}
++}
++
++@article{collins1994,
++  author = {Collins, Daniel W. and Kothari, S.P. and Shanken, Jay and Sloan, Richard G.},
++  title = {{Lack of timeliness and noise as explanations for the low contemporaneuos return-earnings association}},
++  journal = {Journal of Accounting and Economics},
++  year = {1994},
++  volume = {18},
++  number = {3},
++  pages = {289--324},
++  doi = {10.1016/0165-4101(94)90024-8}
++}
++
++@article{foster1977,
++  author = {Foster, George},
++  title = {{Quarterly Accounting Data: Time-Series Properties and Predicative-Ability Results.}},
++  journal = {The Accounting Review},
++  year = {1977},
++  volume = {52},
++  number = {1},
++  pages = {1--21},
++  doi = {10.2308/tar-4493420}
++}
++
++@article{weller2018,
++  author = {Weller, Brian M.},
++  title = {{Does Algorithmic Trading Reduce Information Acquisition?}},
++  journal = {The Review of Financial Studies},
++  year = {2018},
++  volume = {31},
++  number = {6},
++  pages = {2184--2226},
++  doi = {10.1093/rfs/hhx137}
++}
++
++@article{bollerslev2018volume,
++  author = {Bollerslev, Tim and Li, Jia and Xue, Yuan},
++  title = {{Volume, Volatility, and Public News Announcements}},
++  journal = {The Review of Economic Studies},
++  year = {2018},
++  volume = {85},
++  number = {4},
++  pages = {2005--2041},
++  doi = {10.1093/restud/rdy003}
++}
++
++@article{patton2012,
++  author = {Patton, Andrew J. and Verardo, Michela},
++  title = {{Does Beta Move with News? Firm-Specific Information Flows and Learning about Profitability}},
++  journal = {Review of Financial Studies},
++  year = {2012},
++  volume = {25},
++  number = {9},
++  pages = {2789--2839},
++  doi = {10.1093/rfs/hhs073}
++}
++
++@article{kaniel2012,
++  author = {KANIEL, RON and LIU, SHUMING and SAAR, GIDEON and TITMAN, SHERIDAN},
++  title = {{Individual Investor Trading and Return Patterns around Earnings Announcements}},
++  journal = {The Journal of Finance},
++  year = {2012},
++  volume = {67},
++  number = {2},
++  pages = {639--680},
++  doi = {10.1111/j.1540-6261.2012.01727.x}
++}
++
++@article{christophe2004,
++  author = {Christophe, Stephen E. and Ferri, Michael G. and Angel, James J.},
++  title = {{Short-Selling Prior to Earnings Announcements}},
++  journal = {The Journal of Finance},
++  year = {2004},
++  volume = {59},
++  number = {4},
++  pages = {1845--1876},
++  doi = {10.1111/j.1540-6261.2004.00681.x}
++}
++
++@article{berkman2009,
++  author = {Berkman, Henk and Dimitrov, Valentin and Jain, Prem C. and Koch, Paul D. and Tice, Sheri},
++  title = {{Sell on the news: Differences of opinion, short-sales constraints, and returns around earnings announcements☆}},
++  journal = {Journal of Financial Economics},
++  year = {2009},
++  volume = {92},
++  number = {3},
++  pages = {376--399},
++  doi = {10.1016/j.jfineco.2008.04.009}
++}
++
++@article{hendershott2015,
++  author = {Hendershott, Terrence and Livdan, Dmitry and Schürhoff, Norman},
++  title = {{Are institutions informed about news?}},
++  journal = {Journal of Financial Economics},
++  year = {2015},
++  volume = {117},
++  number = {2},
++  pages = {249--287},
++  doi = {10.1016/j.jfineco.2015.03.007}
++}
++
++@article{lee2012,
++  author = {Lee, Suzanne S.},
++  title = {{Jumps and Information Flow in Financial Markets}},
++  journal = {Review of Financial Studies},
++  year = {2012},
++  volume = {25},
++  number = {2},
++  pages = {439--479},
++  doi = {10.1093/rfs/hhr084}
++}
++
++@article{cao2024,
++  author = {Cao, Sean and Jiang, Wei and Wang, Junbo and Yang, Baozhong},
++  title = {{From Man vs. Machine to Man + Machine: The art and AI of stock analyses}},
++  journal = {Journal of Financial Economics},
++  year = {2024},
++  volume = {160},
++  pages = {103910},
++  doi = {10.1016/j.jfineco.2024.103910}
++}
++
++@article{bali2023,
++  author = {Bali, Turan G and Beckmeyer, Heiner and Mörke, Mathis and Weigert, Florian},
++  title = {{Option Return Predictability with Machine Learning and Big Data}},
++  journal = {The Review of Financial Studies},
++  year = {2023},
++  volume = {36},
++  number = {9},
++  pages = {3548--3602},
++  doi = {10.1093/rfs/hhad017}
++}
++
++@article{easton2024,
++  author = {Easton, Peter D. and Kapons, Martin M. and Monahan, Steven J. and Schütt, Harm H. and Weisbrod, Eric H.},
++  title = {{Forecasting Earnings Using \emph{k}-Nearest Neighbors}},
++  journal = {The Accounting Review},
++  year = {2024},
++  volume = {99},
++  number = {3},
++  pages = {115--140},
++  doi = {10.2308/tar-2021-0478}
++}
++
++@article{so2013,
++  author = {So, Eric C.},
++  title = {{A new approach to predicting analyst forecast errors: Do investors overweight analyst forecasts?}},
++  journal = {Journal of Financial Economics},
++  year = {2013},
++  volume = {108},
++  number = {3},
++  pages = {615--640},
++  doi = {10.1016/j.jfineco.2013.02.002}
++}
++
++@article{hou2012,
++  author = {Hou, Kewei and van Dijk, Mathijs A. and Zhang, Yinglei},
++  title = {{The implied cost of capital: A new approach}},
++  journal = {Journal of Accounting and Economics},
++  year = {2012},
++  volume = {53},
++  number = {3},
++  pages = {504--526},
++  doi = {10.1016/j.jacceco.2011.12.001}
++}
++
++@article{hansen2006,
++  author = {Hansen, Peter R and Lunde, Asger},
++  title = {{Realized Variance and Market Microstructure Noise}},
++  journal = {Journal of Business \& Economic Statistics},
++  year = {2006},
++  volume = {24},
++  number = {2},
++  pages = {127--161},
++  doi = {10.1198/073500106000000071}
++}
++
++@article{barndorff2008,
++  author = {Barndorff-Nielsen, Ole E. and Hansen, Peter Reinhard and Lunde, Asger and Shephard, Neil},
++  title = {{Designing Realized Kernels to Measure the ex post Variation of Equity Prices in the Presence of Noise}},
++  journal = {Econometrica},
++  year = {2008},
++  volume = {76},
++  number = {6},
++  pages = {1481--1536},
++  doi = {10.3982/ecta6495}
++}
++
++@article{vankervel2019,
++  author = {VAN KERVEL, VINCENT and MENKVELD, ALBERT J.},
++  title = {{High-Frequency Trading around Large Institutional Orders}},
++  journal = {The Journal of Finance},
++  year = {2019},
++  volume = {74},
++  number = {3},
++  pages = {1091--1137},
++  doi = {10.1111/jofi.12759}
++}
++
++@article{avramov2023,
++  author = {Avramov, Doron and Cheng, Si and Metzker, Lior},
++  title = {{Machine Learning vs. Economic Restrictions: Evidence from Stock Return Predictability}},
++  journal = {Management Science},
++  year = {2023},
++  volume = {69},
++  number = {5},
++  pages = {2587--2619},
++  doi = {10.1287/mnsc.2022.4449}
++}
++
++@article{chordia2020,
++  author = {Chordia, Tarun and Goyal, Amit and Saretto, Alessio},
++  title = {{Anomalies and False Rejections}},
++  journal = {The Review of Financial Studies},
++  year = {2020},
++  volume = {33},
++  number = {5},
++  pages = {2134--2179},
++  doi = {10.1093/rfs/hhaa018}
++}
++
++@article{barber2013,
++  author = {Barber, Brad M. and De George, Emmanuel T. and Lehavy, Reuven and Trueman, Brett},
++  title = {{The earnings announcement premium around the globe}},
++  journal = {Journal of Financial Economics},
++  year = {2013},
++  volume = {108},
++  number = {1},
++  pages = {118--138},
++  doi = {10.1016/j.jfineco.2012.10.006}
++}
++
++@article{politis1994,
++  author = {Politis, Dimitris N. and Romano, Joseph P.},
++  title = {{The Stationary Bootstrap}},
++  journal = {Journal of the American Statistical Association},
++  year = {1994},
++  volume = {89},
++  number = {428},
++  pages = {1303--1313},
++  doi = {10.1080/01621459.1994.10476870}
++}
++
++@article{hansen2011mcs,
++  author = {Hansen, Peter R. and Lunde, Asger and Nason, James M.},
++  title = {{The Model Confidence Set}},
++  journal = {Econometrica},
++  year = {2011},
++  volume = {79},
++  number = {2},
++  pages = {453--497},
++  doi = {10.3982/ecta5771}
++}
++
++@article{benjamini1995,
++  author = {Benjamini, Yoav and Hochberg, Yosef},
++  title = {{Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing}},
++  journal = {Journal of the Royal Statistical Society Series B: Statistical Methodology},
++  year = {1995},
++  volume = {57},
++  number = {1},
++  pages = {289--300},
++  doi = {10.1111/j.2517-6161.1995.tb02031.x}
++}
++
++@article{amihud2002,
++  author = {Amihud, Yakov},
++  title = {{Illiquidity and stock returns: cross-section and time-series effects}},
++  journal = {Journal of Financial Markets},
++  year = {2002},
++  volume = {5},
++  number = {1},
++  pages = {31--56},
++  doi = {10.1016/s1386-4181(01)00024-6}
++}
++
++@article{fama1993,
++  author = {Fama, Eugene F. and French, Kenneth R.},
++  title = {{Common risk factors in the returns on stocks and bonds}},
++  journal = {Journal of Financial Economics},
++  year = {1993},
++  volume = {33},
++  number = {1},
++  pages = {3--56},
++  doi = {10.1016/0304-405x(93)90023-5}
++}
++
++@article{grossman1980,
++  author = {Grossman, Sanford J. and Stiglitz, Joseph E.},
++  title = {{On the Impossibility of Informationally Efficient Markets}},
++  journal = {American Economic Review},
++  year = {1980},
++  volume = {70},
++  number = {3},
++  pages = {393--408}
++}
++@article{holm1979,
++  author = {Holm, Sture},
++  title = {{A Simple Sequentially Rejective Multiple Test Procedure}},
++  journal = {Scandinavian Journal of Statistics},
++  year = {1979},
++  volume = {6},
++  number = {2},
++  pages = {65--70}
++}
++@article{kunsch1989,
++  author = {K{\"u}nsch, Hans R.},
++  title = {{The Jackknife and the Bootstrap for General Stationary Observations}},
++  journal = {Annals of Statistics},
++  year = {1989},
++  volume = {17},
++  number = {3},
++  pages = {1217--1241},
++  doi = {10.1214/aos/1176347265}
++}
++@inproceedings{ke2017lightgbm,
++  author = {Ke, Guolin and Meng, Qi and Finley, Thomas and Wang, Taifeng and Chen, Wei and Ma, Weidong and Ye, Qiwei and Liu, Tie-Yan},
++  title = {{LightGBM: A Highly Efficient Gradient Boosting Decision Tree}},
++  booktitle = {Advances in Neural Information Processing Systems 30},
++  year = {2017},
++  pages = {3146--3154}
++}
++@book{lopezdeprado2018,
++  author = {L{\'o}pez de Prado, Marcos},
++  title = {{Advances in Financial Machine Learning}},
++  publisher = {Wiley},
++  address = {Hoboken, NJ},
++  year = {2018}
++}
++@techreport{frazzini2018costs,
++  author = {Frazzini, Andrea and Israel, Ronen and Moskowitz, Tobias J.},
++  title = {{Trading Costs}},
++  institution = {SSRN Working Paper 3229719},
++  year = {2018},
++  doi = {10.2139/ssrn.3229719}
++}
++@incollection{barndorff2010semivariance,
++  author = {Barndorff-Nielsen, Ole E. and Kinnebrock, Silja and Shephard, Neil},
++  title = {{Measuring Downside Risk: Realised Semivariance}},
++  booktitle = {Volatility and Time Series Econometrics: Essays in Honor of Robert F. Engle},
++  editor = {Bollerslev, Tim and Russell, Jeffrey and Watson, Mark},
++  publisher = {Oxford University Press},
++  year = {2010},
++  pages = {117--136},
++  doi = {10.1093/acprof:oso/9780199549498.003.0007}
++}
++@article{boucher2026wp15,
++  author = {Boucher, Simon-Pierre},
++  title = {{Faster, Wider, More Fragile? The Evolution of Intraday Shock Propagation in U.S. Equity Markets, 1992--2026}},
++  journal = {UQO Working Paper No. 15},
++  year = {2026}
++}
++
++@article{harvey2016,
++  author = {Harvey, Campbell R. and Liu, Yan and Zhu, Heqing},
++  title = {{\ldots and the Cross-Section of Expected Returns}},
++  journal = {The Review of Financial Studies},
++  year = {2016},
++  volume = {29},
++  number = {1},
++  pages = {5--68},
++  doi = {10.1093/rfs/hhv059}
++}

added paper/.latexmkrc +5 −0

Large diff (5 lines) is not loaded by default to keep the page fast.Load diff

added paper/Makefile +11 −0

Large diff (11 lines) is not loaded by default to keep the page fast.Load diff

added paper/appendix/appendix.tex +54 −0

Large diff (54 lines) is not loaded by default to keep the page fast.Load diff

added paper/main.tex +141 −0

Large diff (141 lines) is not loaded by default to keep the page fast.Load diff

added paper/references.bib +1021 −0

Large diff (1,021 lines) is not loaded by default to keep the page fast.Load diff

added paper/sections/data.tex +83 −0

Large diff (83 lines) is not loaded by default to keep the page fast.Load diff

added paper/sections/literature.tex +46 −0

Large diff (46 lines) is not loaded by default to keep the page fast.Load diff

added paper/sections/methodology.tex +78 −0

Large diff (78 lines) is not loaded by default to keep the page fast.Load diff

added paper/sections/theory_hypotheses.tex +64 −0

Large diff (64 lines) is not loaded by default to keep the page fast.Load diff

added paper/sections/titlepage.tex +77 −0

Large diff (77 lines) is not loaded by default to keep the page fast.Load diff

added paper/sections/titlepage_anon.tex +53 −0

Large diff (53 lines) is not loaded by default to keep the page fast.Load diff

added paper/uq_logo.jpg +0 −0

Binary file not shown.

added pyproject.toml +42 −0

Large diff (42 lines) is not loaded by default to keep the page fast.Load diff

added scripts/00_probe_findata.py +42 −0

Large diff (42 lines) is not loaded by default to keep the page fast.Load diff

added scripts/01_audit_universe.py +83 −0

Large diff (83 lines) is not loaded by default to keep the page fast.Load diff

added scripts/01a_extract_node_tables.sh +16 −0

Large diff (16 lines) is not loaded by default to keep the page fast.Load diff

added scripts/02_build_pit_panel.py +75 −0

Large diff (75 lines) is not loaded by default to keep the page fast.Load diff

added scripts/03_build_features.py +116 −0

Large diff (116 lines) is not loaded by default to keep the page fast.Load diff

added scripts/04_build_labels.py +83 −0

Large diff (83 lines) is not loaded by default to keep the page fast.Load diff

added scripts/05_train_baselines.py +108 −0

Large diff (108 lines) is not loaded by default to keep the page fast.Load diff

added scripts/06_train_advanced.py +116 −0

Large diff (116 lines) is not loaded by default to keep the page fast.Load diff

added scripts/07_evaluate.py +209 −0

Large diff (209 lines) is not loaded by default to keep the page fast.Load diff

added scripts/08_run_robustness.py +156 −0

Large diff (156 lines) is not loaded by default to keep the page fast.Load diff

added scripts/09_make_manuscript.py +203 −0

Large diff (203 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/__init__.py +2 −0

Large diff (2 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/api_client.py +226 −0

Large diff (226 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/calendar.py +78 −0

Large diff (78 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/catalog.py +31 −0

Large diff (31 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/corporate_actions.py +36 −0

Large diff (36 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/features.py +181 −0

Large diff (181 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/figures.py +156 −0

Large diff (156 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/inference.py +134 −0

Large diff (134 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/ingest.py +63 −0

Large diff (63 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/instruments.py +119 −0

Large diff (119 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/labels.py +104 −0

Large diff (104 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/models.py +137 −0

Large diff (137 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/pipeline.py +62 −0

Large diff (62 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/pit_fundamentals.py +115 −0

Large diff (115 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/portfolio.py +124 −0

Large diff (124 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/provenance.py +61 −0

Large diff (61 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/robustness.py +46 −0

Large diff (46 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/samples.py +180 −0

Large diff (180 lines) is not loaded by default to keep the page fast.Load diff

added src/wp39/splits.py +65 −0

Large diff (65 lines) is not loaded by default to keep the page fast.Load diff

added tests/__init__.py +0 −0

Large diff (0 lines) is not loaded by default to keep the page fast.Load diff

added tests/conftest.py +39 −0

Large diff (39 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_api_contract.py +42 −0

Large diff (42 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_corporate_actions.py +22 −0

Large diff (22 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_features.py +37 −0

Large diff (37 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_instrument_ids.py +17 −0

Large diff (17 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_labels.py +26 −0

Large diff (26 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_pit_leakage.py +45 −0

Large diff (45 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_reproducibility.py +23 −0

Large diff (23 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_splits.py +37 −0

Large diff (37 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_timestamp_dsts.py +30 −0

Large diff (30 lines) is not loaded by default to keep the page fast.Load diff

added tests/test_train_only_transforms.py +15 −0

Large diff (16 lines) is not loaded by default to keep the page fast.Load diff

spb/wp39_uqo

WP39: phase 0-2 scaffold — data audit, frozen design, PIT protocol, client, features, tests

Theme