1{"text": "{\"genre\":\"empirical\",\"subfield\":\"asset_pricing\",\"subfield_secondary\":\"behavioral_finance\",\"primary_method\":\"reduced_form_regression\",\"identification\":{\"estimand\":\"Association between past factor-related returns (and the flows they attract) and funds' subsequent alphas, holding AUM or factor-adjusted expected returns fixed\",\"strategy\":\"Annual conditional portfolio sorts: funds are sorted on AUM or seven-factor alpha, then split into terciles of past factor-related returns, and future benchmark-adjusted returns are compared. Cross-sectional variation in trading costs and style-level flow-induced trading is used to support a scale-effect mechanism.\",\"key_assumptions\":[\"Factor-related returns reflect no managerial skill, so flows they attract are uninformative\",\"The seven-factor model adequately separates factor exposure from skill\",\"Diminishing returns to scale operate through trading costs\"],\"threats_addressed\":[\"Differences in prior real alpha\",\"Differences in fund size alone\",\"Crisis periods and extreme past returns\",\"Choice of factor model and estimation window\",\"Cap-based vs style-based fund classification\",\"Style-level crowding\"],\"causal_language_strength\":3,\"uses_causal_claim\":true},\"abstract_moves\":{\"question\":{\"present\":false,\"sentence_index\":-1},\"gap\":{\"present\":false,\"sentence_index\":-1},\"method\":{\"present\":false,\"sentence_index\":-1},\"results\":{\"present\":true,\"sentence_index\":0},\"magnitude\":{\"present\":false,\"sentence_index\":-1},\"contribution\":{\"present\":true,\"sentence_index\":3}},\"introduction\":{\"moves\":[{\"paragraph_index\":0,\"move\":\"motivation\"},{\"paragraph_index\":0,\"move\":\"tension_or_puzzle\"},{\"paragraph_index\":0,\"move\":\"mechanism_or_theory_sketch\"},{\"paragraph_index\":1,\"move\":\"preview_of_approach\"},{\"paragraph_index\":1,\"move\":\"robustness_mention\"},{\"paragraph_index\":2,\"move\":\"preview_of_results\"},{\"paragraph_index\":3,\"move\":\"preview_of_results\"},{\"paragraph_index\":4,\"move\":\"preview_of_results\"},{\"paragraph_index\":4,\"move\":\"identification_defense\"},{\"paragraph_index\":4,\"move\":\"robustness_mention\"},{\"paragraph_index\":5,\"move\":\"contribution_statement\"},{\"paragraph_index\":6,\"move\":\"related_literature_positioning\"},{\"paragraph_index\":7,\"move\":\"roadmap\"}],\"contribution_paragraph_index\":5,\"contribution_types\":[\"new_evidence\",\"resolution_of_puzzle\"],\"roadmap_present\":true,\"literature_positioning_location\":\"inside_introduction\"},\"section_roles\":[{\"idx\":0,\"role\":\"introduction\"},{\"idx\":1,\"role\":\"data\"},{\"idx\":2,\"role\":\"results\"},{\"idx\":3,\"role\":\"results\"},{\"idx\":4,\"role\":\"results\"},{\"idx\":5,\"role\":\"results\"},{\"idx\":6,\"role\":\"conclusion\"},{\"idx\":7,\"role\":\"appendix\"},{\"idx\":8,\"role\":\"appendix\"},{\"idx\":9,\"role\":\"appendix\"},{\"idx\":10,\"role\":\"appendix\"},{\"idx\":11,\"role\":\"appendix\"},{\"idx\":12,\"role\":\"appendix\"},{\"idx\":13,\"role\":\"appendix\"}],\"results_presentation\":{\"economic_significance_discussed\":true,\"statistical_significance_language\":\"unclear\",\"magnitude_statements_in_intro\":true},\"robustness\":{\"location\":\"appendix\",\"types\":[\"alternative_specification\",\"subsample\",\"alternative_measure\",\"alternative_sample_period\"]},\"tables\":{\"convention_notes\":\"Table notes open with the sorting procedure ('For each calendar year of data...'), then describe the conditional double sorts into terciles or quintiles, the estimation windows, and AUM-weighted or equal-weighted annualized returns. Most tables are portfolio-sort tables; some report time-series regressions on style returns.\",\"units_stated_in_notes\":true,\"standard_errors_description\":\"unclear\"},\"figures\":{\"uncertainty_shown\":\"unclear\"},\"conclusion\":{\"bounded_claims\":true,\"policy_implications\":false,\"open_questions\":false,\"restates_contribution\":true},\"hedging\":{\"level\":\"moderate\",\"typical_devices\":[\"'consistent with' framing\",\"'likely due to' for the mechanism\",\"'in this sense' qualifiers\",\"'can be explained by'\"]},\"claim_evidence\":[{\"claim\":\"Fund flows rise with factor-related returns after controlling for factor-adjusted expected returns\",\"evidence_type\":\"table\",\"evidence_ref\":\"Table I\"},{\"claim\":\"Within AUM quintiles, top-tercile FRR funds underperform bottom-tercile FRR funds by about 300-400 bps per year\",\"evidence_type\":\"table\",\"evidence_ref\":\"Table II\"},{\"claim\":\"The underperformance persists after controlling for prior seven-factor alpha\",\"evidence_type\":\"table\",\"evidence_ref\":\"Table III\"},{\"claim\":\"Underperformance of high-FRR funds is concentrated among funds with high trading costs\",\"evidence_type\":\"table\",\"evidence_ref\":\"Table IV\"},{\"claim\":\"Styles with high flow-induced trading earn lower future returns\",\"evidence_type\":\"table\",\"evidence_ref\":\"Table V, Table BIII\"},{\"claim\":\"Negative aggregate active-fund performance is driven mainly by high-FRR funds; investors lose about $15bn per year\",\"evidence_type\":\"table\",\"evidence_ref\":\"Table VII\"},{\"claim\":\"Results are robust to alternative windows, factor models, subsamples, and controlling for AUM and alpha jointly\",\"evidence_type\":\"table\",\"evidence_ref\":\"Tables AI-AIII, Figure A1\"}],\"structural_summary\":\"The introduction opens with a puzzle: Berk–Green predicts zero net alpha, yet aggregate alpha is negative. It challenges BG's assumption of rational investors and previews the data and factor model. It then walks through the results sequentially: flows respond to factor returns, high-FRR funds underperform, the mechanism runs through trading costs, and aggregate performance follows. It closes with the contribution, contrast with a contemporaneous study, and a roadmap. The body moves from data, to flows, to future alphas, to mechanism (fund-level scale and style crowding), to aggregate implications. A brief conclusion restates the findings, and appendices hold the robustness checks.\",\"writing_profile_keywords\":[\"portfolio sorts\",\"mutual fund flows\",\"decreasing returns to scale\",\"sequential results narrative\",\"first-person singular\",\"magnitudes in bps\",\"appendix robustness\"],\"confidence\":0.82,\"notes\":\"Single-author paper written in the first person. The causal claim about flows causing underperformance is framed as 'consistent with' and 'likely'. Statistical significance conventions (stars, t-stats, p-values) are not visible in the excerpts.\"}", "json": {"genre": "empirical", "subfield": "asset_pricing", "subfield_secondary": "behavioral_finance", "primary_method": "reduced_form_regression", "identification": {"estimand": "Association between past factor-related returns (and the flows they attract) and funds' subsequent alphas, holding AUM or factor-adjusted expected returns fixed", "strategy": "Annual conditional portfolio sorts: funds are sorted on AUM or seven-factor alpha, then split into terciles of past factor-related returns, and future benchmark-adjusted returns are compared. Cross-sectional variation in trading costs and style-level flow-induced trading is used to support a scale-effect mechanism.", "key_assumptions": ["Factor-related returns reflect no managerial skill, so flows they attract are uninformative", "The seven-factor model adequately separates factor exposure from skill", "Diminishing returns to scale operate through trading costs"], "threats_addressed": ["Differences in prior real alpha", "Differences in fund size alone", "Crisis periods and extreme past returns", "Choice of factor model and estimation window", "Cap-based vs style-based fund classification", "Style-level crowding"], "causal_language_strength": 3, "uses_causal_claim": true}, "abstract_moves": {"question": {"present": false, "sentence_index": -1}, "gap": {"present": false, "sentence_index": -1}, "method": {"present": false, "sentence_index": -1}, "results": {"present": true, "sentence_index": 0}, "magnitude": {"present": false, "sentence_index": -1}, "contribution": {"present": true, "sentence_index": 3}}, "introduction": {"moves": [{"paragraph_index": 0, "move": "motivation"}, {"paragraph_index": 0, "move": "tension_or_puzzle"}, {"paragraph_index": 0, "move": "mechanism_or_theory_sketch"}, {"paragraph_index": 1, "move": "preview_of_approach"}, {"paragraph_index": 1, "move": "robustness_mention"}, {"paragraph_index": 2, "move": "preview_of_results"}, {"paragraph_index": 3, "move": "preview_of_results"}, {"paragraph_index": 4, "move": "preview_of_results"}, {"paragraph_index": 4, "move": "identification_defense"}, {"paragraph_index": 4, "move": "robustness_mention"}, {"paragraph_index": 5, "move": "contribution_statement"}, {"paragraph_index": 6, "move": "related_literature_positioning"}, {"paragraph_index": 7, "move": "roadmap"}], "contribution_paragraph_index": 5, "contribution_types": ["new_evidence", "resolution_of_puzzle"], "roadmap_present": true, "literature_positioning_location": "inside_introduction"}, "section_roles": [{"idx": 0, "role": "introduction"}, {"idx": 1, "role": "data"}, {"idx": 2, "role": "results"}, {"idx": 3, "role": "results"}, {"idx": 4, "role": "results"}, {"idx": 5, "role": "results"}, {"idx": 6, "role": "conclusion"}, {"idx": 7, "role": "appendix"}, {"idx": 8, "role": "appendix"}, {"idx": 9, "role": "appendix"}, {"idx": 10, "role": "appendix"}, {"idx": 11, "role": "appendix"}, {"idx": 12, "role": "appendix"}, {"idx": 13, "role": "appendix"}], "results_presentation": {"economic_significance_discussed": true, "statistical_significance_language": "unclear", "magnitude_statements_in_intro": true}, "robustness": {"location": "appendix", "types": ["alternative_specification", "subsample", "alternative_measure", "alternative_sample_period"]}, "tables": {"convention_notes": "Table notes open with the sorting procedure ('For each calendar year of data...'), then describe the conditional double sorts into terciles or quintiles, the estimation windows, and AUM-weighted or equal-weighted annualized returns. Most tables are portfolio-sort tables; some report time-series regressions on style returns.", "units_stated_in_notes": true, "standard_errors_description": "unclear"}, "figures": {"uncertainty_shown": "unclear"}, "conclusion": {"bounded_claims": true, "policy_implications": false, "open_questions": false, "restates_contribution": true}, "hedging": {"level": "moderate", "typical_devices": ["'consistent with' framing", "'likely due to' for the mechanism", "'in this sense' qualifiers", "'can be explained by'"]}, "claim_evidence": [{"claim": "Fund flows rise with factor-related returns after controlling for factor-adjusted expected returns", "evidence_type": "table", "evidence_ref": "Table I"}, {"claim": "Within AUM quintiles, top-tercile FRR funds underperform bottom-tercile FRR funds by about 300-400 bps per year", "evidence_type": "table", "evidence_ref": "Table II"}, {"claim": "The underperformance persists after controlling for prior seven-factor alpha", "evidence_type": "table", "evidence_ref": "Table III"}, {"claim": "Underperformance of high-FRR funds is concentrated among funds with high trading costs", "evidence_type": "table", "evidence_ref": "Table IV"}, {"claim": "Styles with high flow-induced trading earn lower future returns", "evidence_type": "table", "evidence_ref": "Table V, Table BIII"}, {"claim": "Negative aggregate active-fund performance is driven mainly by high-FRR funds; investors lose about $15bn per year", "evidence_type": "table", "evidence_ref": "Table VII"}, {"claim": "Results are robust to alternative windows, factor models, subsamples, and controlling for AUM and alpha jointly", "evidence_type": "table", "evidence_ref": "Tables AI-AIII, Figure A1"}], "structural_summary": "The introduction opens with a puzzle: Berk–Green predicts zero net alpha, yet aggregate alpha is negative. It challenges BG's assumption of rational investors and previews the data and factor model. It then walks through the results sequentially: flows respond to factor returns, high-FRR funds underperform, the mechanism runs through trading costs, and aggregate performance follows. It closes with the contribution, contrast with a contemporaneous study, and a roadmap. The body moves from data, to flows, to future alphas, to mechanism (fund-level scale and style crowding), to aggregate implications. A brief conclusion restates the findings, and appendices hold the robustness checks.", "writing_profile_keywords": ["portfolio sorts", "mutual fund flows", "decreasing returns to scale", "sequential results narrative", "first-person singular", "magnitudes in bps", "appendix robustness"], "confidence": 0.82, "notes": "Single-author paper written in the first person. The causal claim about flows causing underperformance is framed as 'consistent with' and 'likely'. Statistical significance conventions (stars, t-stats, p-values) are not visible in the excerpts."}, "model": "claude-opus-5-5", "stop_reason": "end_turn", "latency_s": 18.68, "usage": {"input": 8520, "output": 2362, "cache_read": 3968, "cache_write": 0}, "cost_usd": 0.082114, "request_id": "req_011CggqwuLnwHm462RUKEHnj", "cached": false, "ts": "2026-11-03T23:15:14+00:00", "effort": "medium"}