research: S&P 600 replication of mid-cap quality sleeves (all REJECT)

Preregistered replication vs IJR, 2017-01..2026-09, 20bp decides:
quality composite +2.3%/yr (t 0.78; halves +8.0/-3.5), quality-filtered
momentum +2.2% (t 0.73); SUE and lottery-filtered momentum also REJECT,
all weaker at 40bp. Equal-weight members beat IJR (t 3.2) and every
sleeve. Decision: drop quality stock selection in mid/small caps.

Adds --universe small to fetch_us_fundamentals (IJR, name-checked CIKs),
a `which` argument to fundamentals.load_mid, and --universe small to
midcap_sleeve.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

CLAUDE.md

@@ -82,7 +82,8 @@ uv run python -m research.lab.fetch_wiki_attention && PYTHONPATH=. uv run --with
 PYTHONPATH=. uv run --with numba python -m research.lab.concentration     # industry caps on the momentum sleeve
 PYTHONPATH=. uv run python -m research.lab.universe_midcap              # overnight_recovery_top10 on S&P 500 ∪ S&P 400 (preregistered; FAIL 2026-10-08)
 uv run python -m research.lab.fetch_us_fundamentals --universe mid      # EDGAR PIT facts for S&P 400 ever-members (name-checked CIKs)
-PYTHONPATH=. uv run python -m research.lab.midcap_sleeve [--only h1..h4]  # separate S&P 400 sleeve vs IJH (preregistered; all REJECT 2026-10-08)
+uv run python -m research.lab.fetch_us_fundamentals --universe small    # the same for S&P 600 (IJR)
+PYTHONPATH=. uv run python -m research.lab.midcap_sleeve [--only h1..h4] [--universe small]  # S&P 400 sleeves vs IJH; S&P 600 replication vs IJR (all REJECT 2026-10-08)
 PYTHONPATH=. uv run python -m research.lab.fetch_shortvolume              # FINRA REG SHO daily short volume (~8y archive, resume-safe)
 PYTHONPATH=. uv run python -m research.lab.short_volume [us|midsmall|all] # SVR (short-volume ratio) IC / quintile scan on PIT panels
 PYTHONPATH=. uv run python -m research.lab.fetch_fails                   # SEC fails-to-deliver archives (2018+, resume-safe)

research/lab/fetch_us_fundamentals.py

@@ -5,8 +5,9 @@ history (needed to turn split-adjusted prices back into market caps).
 
     uv run python -m research.lab.fetch_us_fundamentals
     uv run python -m research.lab.fetch_us_fundamentals --universe mid   # S&P 400 (IJH) ever-members
+    uv run python -m research.lab.fetch_us_fundamentals --universe small # S&P 600 (IJR) ever-members
 
-Writes data/research/us_fund_raw.pkl (us_mid_fund_raw.pkl for --universe mid):
+Writes data/research/us_fund_raw.pkl (us_{mid,small}_fund_raw.pkl for --universe):
     {"facts": DataFrame[ticker, tag, start, end, val, filed, form],
      "splits": DataFrame (dates x tickers, split ratio on split days)}
 
@@ -27,7 +28,7 @@ import yfinance as yf
 from research.lab.fetch_us import _proxy_env
 
 OUT = os.path.join("data", "research", "us_fund_raw.pkl")
-MID_OUT = os.path.join("data", "research", "us_mid_fund_raw.pkl")
+UNIVERSE_FUND = {"mid": "IJH", "small": "IJR"}       # --universe -> iShares fund
 UA = {"User-Agent": "quant-research/0.1 gahow@example.com", "Accept-Encoding": "gzip"}
 FORMS = {"10-K", "10-Q", "10-K/A", "10-Q/A", "10-KT", "20-F", "40-F"}
 
@@ -148,8 +149,8 @@ def _name_tokens(name: str) -> set[str]:
     return {w for w in words if w not in _SUFFIXES and len(w) > 1}
 
 
-def mid_ciks(tickers: list[str]) -> tuple[dict[str, int], pd.DataFrame]:
+def mid_ciks(tickers: list[str], fund: str = "IJH") -> tuple[dict[str, int], pd.DataFrame]:
-    """CIKs for S&P 400 tickers. Candidates come from sec_sic.csv and
+    """CIKs for S&P 400 (IJH) or S&P 600 (IJR) tickers. Candidates come from sec_sic.csv and
     mid_sec_sic.csv (earlier runs), then today's company_tickers.json; both
     map a ticker to whoever holds it NOW, so a candidate counts only if one of
     its SEC names (current or former) shares a word with an iShares holding
@@ -177,7 +178,7 @@ def mid_ciks(tickers: list[str]) -> tuple[dict[str, int], pd.DataFrame]:
                 js = r.json()
                 names[str(c)] = [js.get("name", "")] + [x.get("name", "") for x in js.get("formerNames", [])]
     hold = load_holdings()
-    hnames = hold[hold["fund"] == "IJH"].groupby("ticker")["name"].apply(lambda s: set(s.dropna()))
+    hnames = hold[hold["fund"] == fund].groupby("ticker")["name"].apply(lambda s: set(s.dropna()))
     rows, ok = [], {}
     for t in tickers:
         c = cand.get(t, 0)
@@ -201,13 +202,14 @@ def mid_ciks(tickers: list[str]) -> tuple[dict[str, int], pd.DataFrame]:
 def main() -> None:
     import sys
     global OUT
-    if "--universe" in sys.argv and sys.argv[sys.argv.index("--universe") + 1] == "mid":
+    uni = sys.argv[sys.argv.index("--universe") + 1] if "--universe" in sys.argv else None
+    if uni in UNIVERSE_FUND:
         from research.lab.panel_mid import load_us_mid
-        OUT = MID_OUT
+        OUT = os.path.join("data", "research", f"us_{uni}_fund_raw.pkl")
-        tickers = list(load_us_mid("mid").close.columns)
+        tickers = list(load_us_mid(uni).close.columns)
-        ciks, audit = mid_ciks(tickers)
+        ciks, audit = mid_ciks(tickers, UNIVERSE_FUND[uni])
-        audit.to_csv(os.path.join("data", "research", "us_mid_fund_cik_audit.csv"), index=False)
+        audit.to_csv(os.path.join("data", "research", f"us_{uni}_fund_cik_audit.csv"), index=False)
-        print(f"{len(ciks)}/{len(tickers)} S&P 400 tickers have a name-checked CIK", flush=True)
+        print(f"{len(ciks)}/{len(tickers)} {UNIVERSE_FUND[uni]} tickers have a name-checked CIK", flush=True)
     else:
         with open(os.path.join("data", "research", "us_ohlcv.pkl"), "rb") as fh:
             tickers = [t for t in pickle.load(fh)["close"].columns]

research/lab/fundamentals.py

@@ -294,6 +294,10 @@ def load(p, rebuild: bool = False, raw_path: str = RAW, cache: str = CACHE) -> d
     return F
 
 
-def load_mid(p, rebuild: bool = False) -> dict[str, pd.DataFrame]:
+def load_mid(p, rebuild: bool = False, which: str = "mid") -> dict[str, pd.DataFrame]:
-    """The same factors for the S&P 400 panel (research/lab/panel_mid.py)."""
+    """The same factors for the S&P 400 ("mid") or S&P 600 ("small") panel
-    return load(p, rebuild, MID_RAW, MID_CACHE)
+    (research/lab/panel_mid.py; raw facts from fetch_us_fundamentals --universe)."""
+    if which == "mid":
+        return load(p, rebuild, MID_RAW, MID_CACHE)
+    return load(p, rebuild, os.path.join("data", "research", f"us_{which}_fund_raw.pkl"),
+                os.path.join("data", "research", f"us_{which}_fund_factors.pkl"))

research/lab/midcap_sleeve.py

@@ -4,6 +4,7 @@ alpha over IJH and improve the book more than holding IJH would?
 
     PYTHONPATH=. uv run python -m research.lab.midcap_sleeve --only h4
     PYTHONPATH=. uv run python -m research.lab.midcap_sleeve --only h1 h2 h3   # needs us_mid_fund_raw.pkl
+    PYTHONPATH=. uv run python -m research.lab.midcap_sleeve --universe small   # S&P 600 replication (REPLICATION)
 
 Background (research/lab/universe_midcap.py): ranking S&P 400 names inside
 the large-cap momentum top 10 lost (Sharpe 0.86 vs 0.98, 2019-26); the S&P
@@ -42,11 +43,22 @@ size. Boundary: gp_a covers 55-70% of members (no gross profit for banks,
 REITs), so h2's median screen also drops most financials. EDGAR CIKs are
 name-checked against iShares holding names (723/845 tickers; members
 covered 84-98% a year from 2017).
+
+S&P 600 replication (2026-10-08, REPLICATION written before the S&P 600
+fundamentals were fetched; 963/1112 tickers, members covered 82-99% a year
+from 2017): the quality result does not replicate. h3 alpha vs IJR +2.3%/yr
+(t 0.78), halves +8.0 / -3.5; h2 +2.2% (t 0.73), halves +5.7 / -1.7; h1
++2.7% (t 0.90); h4 +2.2% (t 0.94); all REJECT, and weaker at 40 bp. The
+equal-weight member basket beats IJR (alpha +2.8%, t 3.2, Sharpe 0.60) and
+every sleeve on Sharpe. Per the rule (h3 alpha <= 0 in a half), quality
+stock selection in mid/small caps is dropped. In both universes the
+selection signals earn their alpha in 2017-21 and lose it in 2022-26.
 """
 
 from __future__ import annotations
 
 import argparse
+import dataclasses
 import json
 import os
 import warnings
@@ -55,6 +67,7 @@ import numpy as np
 import pandas as pd
 
 from research.lab.alpha import alpha_stats
+from research.lab.engine import Costs
 from research.lab.factors import overnight_mom, rec_126
 from research.lab.panel_mid import load_us_mid
 from research.lab.us_final import monthly_blend
@@ -96,6 +109,24 @@ PREREG = {
     "reference_only": f"{REFERENCE}: price coverage 57-78%, survivor-biased",
     "status": "holdout overlaps periods seen in earlier mid/small scans; not independent OOS",
 }
+REPLICATION = {
+    "question": "Do the S&P 400 quality sleeves (h3 primary, h2 secondary) replicate on S&P 600?",
+    "written": "2026-10-08, after the S&P 400 results (h3 +3.4%/yr t 1.23, h2 +3.5% t 1.28) and before any "
+               "S&P 600 fundamentals were fetched",
+    "definitions": "unchanged from PREREG (signals, top 30, every 21, buffer 1, 21 phases); universe = PIT S&P 600 "
+                   "members (IJR holdings), benchmark IJR; c0, h1, h4 reported, not judged",
+    "accept_if_all": [
+        "A1: alpha vs IJR over 2017-01-03..2026-09-24, Newey-West t >= 2.0, or t >= 1.5 and alpha > 0 in both halves",
+        "A2: Sharpe(2/3 book + 1/3 sleeve) > Sharpe(2/3 book + 1/3 IJR)",
+        "A3 (h2): Sharpe(h2) > Sharpe(c0_momentum)",
+    ],
+    "costs": "20 bp a side decides; 40 bp a side reported (small-cap spreads)",
+    "decision": {
+        "h3 accepted": "propose a forward-only S&P 400+600 quality sleeve in a validation book",
+        "h3 rejected, alpha > 0 in both halves": "consistent but unproven; no action, no further tuning on 2017-26",
+        "h3 alpha <= 0 in either half": "drop quality stock selection in mid/small caps",
+    },
+}
 
 
 def lottery_ok(p, members: pd.DataFrame) -> pd.DataFrame:
@@ -116,7 +147,7 @@ def score(factors: list[pd.DataFrame], eligible: pd.DataFrame) -> pd.DataFrame:
     return _blend(*[f.where(eligible).rank(axis=1, pct=True) for f in factors]).where(eligible)
 
 
-def sleeves(p, which: list[str]) -> dict[str, pd.DataFrame]:
+def sleeves(p, which: list[str], universe: str = "mid") -> dict[str, pd.DataFrame]:
     mem = p.member
     mom = [overnight_mom(p), rec_126(p)]
     out = {"c0_momentum": score(mom, mem)}
@@ -124,7 +155,7 @@ def sleeves(p, which: list[str]) -> dict[str, pd.DataFrame]:
         out["h4_lottery_excluded_momentum"] = score(mom, mem & lottery_ok(p, mem))
     if {"h1", "h2", "h3"} & set(which):
         from research.lab import fundamentals as fu
-        F = fu.load_mid(p)
+        F = fu.load_mid(p, which=universe)
         if "h1" in which:
             out["h1_earnings_momentum"] = score([F["sue"], F["sue_rev"]], mem)
         if "h2" in which:
@@ -141,6 +172,7 @@ def sharpe(r: pd.Series) -> float:
 
 
 def describe(r: pd.Series, ijh: pd.Series, book: pd.Series) -> dict:
+    """Holdout stats; ``ijh`` is the universe's ETF (IJH or IJR)."""
     h = r.loc[HOLDOUT[0]:HOLDOUT[1]]
     e = (1 + h).cumprod()
     a = alpha_stats(h, ijh)
@@ -172,48 +204,57 @@ def verdict(name: str, d: dict, c0: dict) -> dict:
 def main() -> None:
     ap = argparse.ArgumentParser()
     ap.add_argument("--only", nargs="+", default=["h1", "h2", "h3", "h4"])
-    which = ap.parse_args().only
+    ap.add_argument("--universe", choices=["mid", "small"], default="mid")
-    os.makedirs(OUT, exist_ok=True)
+    args = ap.parse_args()
-    pre = os.path.join(OUT, "prereg.json")
+    which, uni = args.only, args.universe
+    out = OUT if uni == "mid" else OUT + "_sp600"
+    etf = "IJH" if uni == "mid" else "IJR"
+    os.makedirs(out, exist_ok=True)
+    pre = os.path.join(out, "prereg.json")
     if not os.path.exists(pre):
         with open(pre, "w") as fh:
-            json.dump(PREREG, fh, indent=2)
+            json.dump(PREREG if uni == "mid" else {**REPLICATION, "base": PREREG}, fh, indent=2)
-    res_path = os.path.join(OUT, "results.json")
+    res_path = os.path.join(out, "results.json")
     done = json.load(open(res_path)) if os.path.exists(res_path) else {}
     clash = [k for k in done if k != "c0_momentum" and k.split("_")[0] in which]
     if clash:
         raise SystemExit(f"already run: {clash}; this study does not overwrite results")
 
-    p = load_us_mid("mid")
+    p = load_us_mid(uni)
-    ijh = p.extra["etf"]["IJH"].pct_change(fill_method=None)
+    bench = p.extra["etf"][etf].pct_change(fill_method=None)
     book = pd.read_pickle(BOOK)["base"]
-    S = sleeves(p, which)
+    S = sleeves(p, which, uni)
     span = [(REFERENCE[0], HOLDOUT[1])]
+    costs = {"": p.costs} if uni == "mid" else {"": p.costs, "@40bp": Costs(buy=0.004, sell=0.004)}
     rets = {}
-    for name, sc in S.items():
+    for tag, c in costs.items():
-        rets[name] = run(p, sc, span, **KW)["_rets"]
+        pc = dataclasses.replace(p, costs=c)
-        print(f"ran {name}", flush=True)
+        for name, sc in S.items():
+            rets[name + tag] = run(pc, sc, span, **KW)["_rets"]
+            print(f"ran {name}{tag}", flush=True)
     ew = p.close.pct_change(fill_method=None).where(p.member.shift(1, fill_value=False)).mean(axis=1)
     rets["bench_ew_members"] = ew
-    rets["bench_ijh"] = ijh
+    rets[f"bench_{etf.lower()}"] = bench
     for name, r in rets.items():
-        d = describe(r.reindex(p.dates).fillna(0.0), ijh, book)
+        d = describe(r.reindex(p.dates).fillna(0.0), bench, book)
         if name.startswith("h"):
-            d.update(verdict(name, d, done.get("c0_momentum") or describe(rets["c0_momentum"], ijh, book)))
+            ctrl = "c0_momentum" + ("@40bp" if name.endswith("@40bp") else "")
+            d.update(verdict(name, d, done.get(ctrl) or describe(rets[ctrl], bench, book)))
         done[name] = d
-        print(f"{name:32s} cagr {d['cagr']*100:5.1f}% sh {d['sharpe']:.2f} mdd {d['mdd']*100:6.1f}% "
+        print(f"{name:38s} cagr {d['cagr']*100:5.1f}% sh {d['sharpe']:.2f} mdd {d['mdd']*100:6.1f}% "
               f"alpha {d['alpha_ijh']*100:5.1f}% t {d['alpha_t']:.2f} halves "
               f"{[round(x*100, 1) for x in d['alpha_halves']]} corr_book {d['corr_book']:.2f} "
-              f"book+sleeve {d['book_sharpe_with_sleeve']:.2f} vs book+IJH {d['book_sharpe_with_ijh']:.2f} "
+              f"book+sleeve {d['book_sharpe_with_sleeve']:.2f} vs book+{etf} {d['book_sharpe_with_ijh']:.2f} "
               f"ref {d['reference_sharpe']:.2f} {d.get('verdict', '')} {d.get('checks', '')}")
     with open(res_path, "w") as fh:
         json.dump(done, fh, indent=2)
-    pd.DataFrame(rets).to_pickle(os.path.join(OUT, f"returns_{'_'.join(which)}.pkl"))
+    pd.DataFrame(rets).to_pickle(os.path.join(out, f"returns_{'_'.join(which)}.pkl"))
     from research.lab.trials import append_trials
     append_trials(os.path.join("data", "research", "results", "trials_us.jsonl"),
-                  [{"name": f"midcap_sleeve:{k}", **{x: y for x, y in v.items() if x != "years"}}
+                  [{"name": f"midcap_sleeve[{uni}]:{k}", **{x: y for x, y in v.items() if x != "years"}}
                    for k, v in done.items() if k in rets],
-                  {"study": "midcap_sleeve", "role": "preregistered family; holdout partly seen"})
+                  {"study": "midcap_sleeve" if uni == "mid" else "midcap_sleeve_sp600_replication",
+                   "role": "preregistered; holdout partly seen"})
 
 
 if __name__ == "__main__":