1{2 "validation_errors": [],3 "doc_id": "jof_12878",4 "ontology_version": "1.0",5 "prompt_version": "1.0",6 "model": "claude-opus-5-5",7 "effort": "medium",8 "annotated_at": "2026-11-03T23:13:32+00:00",9 "source_sha256": "d1bfcf7d070b759fa3abf18c5abe44a89a780da15d8c1502e6a6b57809856a56",10 "parser": "pymupdf-jof-heuristics-1.3",11 "payload_chars": 28532,12 "usage": {13 "input": 9312,14 "output": 2112,15 "cache_read": 3968,16 "cache_write": 017 },18 "cost_usd": 0.080282,19 "request_id": "req_011CggqpVokERFULLTx3f2Dj",20 "annotation": {21 "genre": "empirical",22 "subfield": "asset_pricing",23 "subfield_secondary": "corporate_finance",24 "primary_method": "reduced_form_regression",25 "identification": {26 "estimand": "Difference in the predictive relation between insider trading strength and future stock returns across short- versus long-horizon insiders.",27 "strategy": "Two-way portfolio sorts on trading strength and horizon, plus panel regressions of future returns on trading strength, horizon and their interaction with controls; supplemented by earnings-surprise, large-price-change and SEC-enforcement tests.",28 "key_assumptions": [29 "Past net order flow proxies for an insider's investment horizon",30 "Horizon-based expectations about continued trading identify unexpected trades",31 "Markets are efficient, so unexpected trades should carry more information"32 ],33 "threats_addressed": [34 "Differences in insider and firm attributes between SH and LH insiders (additional controls)",35 "Overlap with the routine/nonroutine classification of CMP",36 "Firm-level aggregation of the return relation"37 ],38 "causal_language_strength": 2,39 "uses_causal_claim": false40 },41 "abstract_moves": {42 "question": {43 "present": true,44 "sentence_index": 045 },46 "gap": {47 "present": false,48 "sentence_index": -149 },50 "method": {51 "present": false,52 "sentence_index": -153 },54 "results": {55 "present": true,56 "sentence_index": 257 },58 "magnitude": {59 "present": false,60 "sentence_index": -161 },62 "contribution": {63 "present": false,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": 0,79 "move": "preview_of_approach"80 },81 {82 "paragraph_index": 1,83 "move": "mechanism_or_theory_sketch"84 },85 {86 "paragraph_index": 2,87 "move": "mechanism_or_theory_sketch"88 },89 {90 "paragraph_index": 3,91 "move": "mechanism_or_theory_sketch"92 },93 {94 "paragraph_index": 4,95 "move": "mechanism_or_theory_sketch"96 },97 {98 "paragraph_index": 4,99 "move": "preview_of_results"100 },101 {102 "paragraph_index": 5,103 "move": "preview_of_results"104 },105 {106 "paragraph_index": 6,107 "move": "preview_of_results"108 },109 {110 "paragraph_index": 7,111 "move": "related_literature_positioning"112 },113 {114 "paragraph_index": 7,115 "move": "contribution_statement"116 }117 ],118 "contribution_paragraph_index": 7,119 "contribution_types": [120 "new_data_or_measure",121 "new_evidence"122 ],123 "roadmap_present": false,124 "literature_positioning_location": "inside_introduction"125 },126 "section_roles": [127 {128 "idx": 0,129 "role": "introduction"130 },131 {132 "idx": 1,133 "role": "data"134 },135 {136 "idx": 2,137 "role": "results"138 },139 {140 "idx": 3,141 "role": "results"142 },143 {144 "idx": 4,145 "role": "results"146 },147 {148 "idx": 5,149 "role": "results"150 },151 {152 "idx": 6,153 "role": "results"154 },155 {156 "idx": 7,157 "role": "robustness"158 },159 {160 "idx": 8,161 "role": "robustness"162 },163 {164 "idx": 9,165 "role": "conclusion"166 },167 {168 "idx": 10,169 "role": "appendix"170 }171 ],172 "results_presentation": {173 "economic_significance_discussed": false,174 "statistical_significance_language": "unclear",175 "magnitude_statements_in_intro": false176 },177 "robustness": {178 "location": "mixed",179 "types": [180 "alternative_specification",181 "other"182 ]183 },184 "tables": {185 "convention_notes": "Table notes define variables, describe the sorting procedure or write out the full regression equation, and state the dependent variable and sample construction. Tables are organized in panels, e.g., buys vs. sells or all trades vs. subsamples.",186 "units_stated_in_notes": false,187 "standard_errors_description": "unclear"188 },189 "figures": {190 "uncertainty_shown": "no"191 },192 "conclusion": {193 "bounded_claims": true,194 "policy_implications": true,195 "open_questions": true,196 "restates_contribution": true197 },198 "hedging": {199 "level": "moderate",200 "typical_devices": [201 "conjecture framing",202 "'consistent with'",203 "'on average'",204 "'suggests'",205 "'tend to'",206 "'is associated with'"207 ]208 },209 "claim_evidence": [210 {211 "claim": "Short-horizon insiders' trades predict returns better than long-horizon insiders' trades for up to 12 months, on both buys and sells.",212 "evidence_type": "table",213 "evidence_ref": "Tables II, III"214 },215 {216 "claim": "Unexpected insider trades carry more information about future returns.",217 "evidence_type": "table",218 "evidence_ref": "Table IV"219 },220 {221 "claim": "Typical SH trades are more unexpected than LH trades.",222 "evidence_type": "table",223 "evidence_ref": "Table V"224 },225 {226 "claim": "SH insiders and their firms display characteristics associated with short-termism, such as more earnings management and less R&D.",227 "evidence_type": "table",228 "evidence_ref": "Tables VI, VII"229 },230 {231 "claim": "SH insider trades better predict earnings surprises and large price changes.",232 "evidence_type": "table",233 "evidence_ref": "Tables VIII, IX"234 },235 {236 "claim": "SH insiders trade less after intensified SEC enforcement.",237 "evidence_type": "table",238 "evidence_ref": "Table X"239 },240 {241 "claim": "The horizon measure adds information beyond the routine/nonroutine classification.",242 "evidence_type": "table",243 "evidence_ref": "Table XI"244 }245 ],246 "structural_summary": "The paper defines a new insider-horizon measure based on past net order flow and builds a conjecture about expected versus unexpected trades, from which it derives two hypotheses. It then documents baseline return predictability with sorts and regressions. Next, it tests the unexpected-trade mechanism, profiles SH/LH insiders and short-termism, traces sources of the information advantage through earnings and price events, and examines responses to SEC enforcement. It differentiates the measure from the routine/nonroutine concept, runs robustness tests, and closes with practitioner and regulatory implications and future research.",247 "writing_profile_keywords": [248 "new measure",249 "conjecture-driven hypotheses",250 "portfolio sorts",251 "panel regressions",252 "mechanism tests",253 "anecdotal examples",254 "policy implications"255 ],256 "confidence": 0.8,257 "notes": "The abstract appears to merge the start of the introduction. Standard error and significance conventions are not visible in the excerpts. The robustness section lists subsections A and C, so B is missing from the excerpt."258 },259 "deterministic_checks": {260 "roadmap_regex": false,261 "contribution_regex_in_intro": false,262 "magnitude_in_abstract": false,263 "abstract_n_sentences": 5,264 "intro_n_paragraphs": 8,265 "intro_n_words": 1401,266 "causal_words_in_abstract": 0,267 "hedge_words_in_abstract": 2,268 "stars_mentioned": false,269 "clustered_se_mentioned": false,270 "internet_appendix_mentioned": true,271 "section_roles_rule": {272 "0": "introduction",273 "1": "data",274 "2": "results",275 "3": "other",276 "4": "other",277 "5": "other",278 "6": "other",279 "7": "other",280 "8": "robustness",281 "9": "conclusion",282 "10": "appendix"283 }284 },285 "schema_valid": true,286 "code_version": "87c73cd-dirty"287}