1{2 "validation_errors": [],3 "doc_id": "jof_12951",4 "ontology_version": "1.0",5 "prompt_version": "1.0",6 "model": "claude-opus-5-5",7 "effort": "medium",8 "annotated_at": "2026-11-03T23:15:07+00:00",9 "source_sha256": "3ded80a132b52de1af1becf62173be38107ed0fa8539d684a3c44bb05f9cdd7a",10 "parser": "pymupdf-jof-heuristics-1.3",11 "payload_chars": 18273,12 "usage": {13 "input": 6346,14 "output": 1751,15 "cache_read": 3968,16 "cache_write": 017 },18 "cost_usd": 0.061198,19 "request_id": "req_011CggqwgtCH5zGkkKAEoqYZ",20 "annotation": {21 "genre": "methodological_econometric",22 "subfield": "financial_econometrics",23 "subfield_secondary": "asset_pricing",24 "primary_method": "calibration_simulation",25 "identification": {26 "estimand": "Type I and Type II error rates (false discovery rate, miss rate) of multiple-testing procedures applied to fund/strategy selection",27 "strategy": "Double-bootstrap simulation on data adjusted so that a fraction p0 of tests are true (set at in-sample estimates) and the rest have zero alpha",28 "key_assumptions": [29 "p0 acts as a sufficient statistic for the alternative",30 "in-sample estimates are reasonable values for true effects",31 "bootstrap preserves the dependence structure of the data"32 ],33 "threats_addressed": [34 "multiple testing / luck",35 "unknown p0, handled by evaluating performance over a range of p0 values",36 "high-dimensional alternative parameter vector"37 ],38 "causal_language_strength": 1,39 "uses_causal_claim": false40 },41 "abstract_moves": {42 "question": {43 "present": true,44 "sentence_index": 045 },46 "gap": {47 "present": true,48 "sentence_index": 049 },50 "method": {51 "present": true,52 "sentence_index": 253 },54 "results": {55 "present": true,56 "sentence_index": 457 },58 "magnitude": {59 "present": false,60 "sentence_index": -161 },62 "contribution": {63 "present": true,64 "sentence_index": 165 }66 },67 "introduction": {68 "moves": [69 {70 "paragraph_index": 0,71 "move": "motivation"72 },73 {74 "paragraph_index": 0,75 "move": "literature_gap"76 },77 {78 "paragraph_index": 1,79 "move": "tension_or_puzzle"80 },81 {82 "paragraph_index": 2,83 "move": "preview_of_approach"84 },85 {86 "paragraph_index": 2,87 "move": "contribution_statement"88 },89 {90 "paragraph_index": 3,91 "move": "related_literature_positioning"92 },93 {94 "paragraph_index": 4,95 "move": "preview_of_results"96 }97 ],98 "contribution_paragraph_index": 2,99 "contribution_types": [100 "new_method",101 "new_evidence"102 ],103 "roadmap_present": false,104 "literature_positioning_location": "inside_introduction"105 },106 "section_roles": [107 {108 "idx": 0,109 "role": "introduction"110 },111 {112 "idx": 1,113 "role": "theory"114 },115 {116 "idx": 2,117 "role": "results"118 },119 {120 "idx": 3,121 "role": "conclusion"122 },123 {124 "idx": 4,125 "role": "appendix"126 }127 ],128 "results_presentation": {129 "economic_significance_discussed": true,130 "statistical_significance_language": "unclear",131 "magnitude_statements_in_intro": false132 },133 "robustness": {134 "location": "within_results",135 "types": [136 "sensitivity_to_parameters",137 "subsample",138 "other"139 ]140 },141 "tables": {142 "convention_notes": "Table notes describe the simulation design in detail, including bootstrap iteration counts (I, J, M) and the grid of p0 values. Simulated Type I and Type II error rates are tabulated across methods and p0 values. Fund-test tables report results for percentile-based test statistics.",143 "units_stated_in_notes": false,144 "standard_errors_description": "none"145 },146 "figures": {147 "uncertainty_shown": "unclear"148 },149 "conclusion": {150 "bounded_claims": true,151 "policy_implications": false,152 "open_questions": false,153 "restates_contribution": true154 },155 "hedging": {156 "level": "moderate",157 "typical_devices": [158 "may lead us to miss",159 "likely more consistent",160 "conditioning results on assumed p0",161 "we expect our method to be useful"162 ]163 },164 "claim_evidence": [165 {166 "claim": "Existing multiple-testing methods rank in a fairly stable order of performance regardless of p0",167 "evidence_type": "table",168 "evidence_ref": "Tables II, III"169 },170 {171 "claim": "Error rates of Yan and Zheng's application of the Fama-French approach are inconsistent with a widespread-mispricing narrative",172 "evidence_type": "table",173 "evidence_ref": "Table IV"174 },175 {176 "claim": "Fama and French (2010) bootstrap test lacks power to detect outperforming funds",177 "evidence_type": "table",178 "evidence_ref": "Tables VI-VIII"179 },180 {181 "claim": "Optimal t-statistic cutoffs vary with the assumed fraction of true strategies p0",182 "evidence_type": "figure",183 "evidence_ref": "Figure 3"184 }185 ],186 "structural_summary": "Opens by framing manager and factor selection as a two-error problem and noting that current tools ignore Type II error and differential error costs. It then explains why Type II error is hard to compute under multiple tests and proposes a double bootstrap that fixes the fraction of true effects p0. The method is positioned against single-bootstrap predecessors (KTWW, Fama-French). A method section develops the error-rate estimation. A long applications section covers backtested strategies, 18,000 anomalies (critiquing Yan-Zheng), a simulation study, and mutual-fund performance evaluation (critiquing Fama-French). A short conclusion restates the contribution and its expected usefulness.",187 "writing_profile_keywords": [188 "multiple testing",189 "double bootstrap",190 "Type II error",191 "false discovery rate",192 "methodological critique",193 "simulation-based"194 ],195 "confidence": 0.8,196 "notes": "Zero equations are recorded despite the methodological content, so formulas may not have been captured in the extraction. Robustness checks appear to be embedded in the applications (simulation study, subsamples, alternative test statistics) rather than in a separate section."197 },198 "deterministic_checks": {199 "roadmap_regex": false,200 "contribution_regex_in_intro": false,201 "magnitude_in_abstract": true,202 "abstract_n_sentences": 5,203 "intro_n_paragraphs": 5,204 "intro_n_words": 910,205 "causal_words_in_abstract": 0,206 "hedge_words_in_abstract": 0,207 "stars_mentioned": false,208 "clustered_se_mentioned": false,209 "internet_appendix_mentioned": false,210 "section_roles_rule": {211 "0": "introduction",212 "1": "other",213 "2": "other",214 "3": "conclusion",215 "4": "appendix"216 }217 },218 "schema_valid": true,219 "code_version": "87c73cd-dirty"220}