Text-NLP#90tier 1live in productionNew

transcript analyst frustration

cadence: Eventdata: mediumlong onlyshort onlylong short
JAR
2010
J. of Accounting Research
Hollander, Pronk & Roelofsen (2010) *Journal of Accounting Research*: "Does Silence Speak? An Empirical Analysis of Disclosure Choices During Conference Calls." Bowen-Davis-Matsumoto (2002 AR) and Larcker-Zakolyukina (2012 JAR) extend: analyst tone in Q&A questions carries forward-looking information.
Citation only — paper link pending.

What it checks

Measures the tone of the ANALYSTS asking questions, not management's answers. When their questions sound frustrated or skeptical, the stock tends to drift down over the following weeks.

Mechanism

Analyst tone during the Q&A section — the *questions* themselves, not the answers — carries forward-looking information independent of the news in the call. When analysts sound skeptical or frustrated (negative FinBERT tone, sharp questions), 60-day drift skews negative. We z-score per-call mean analyst_qa_tone against the firm's own history of analyst sentiment.

Signal rule

SHORT when analyst_qa_tone own-z <= -1.0 (analysts frustrated). LONG when own-z >= +1.0 (analysts enthusiastic).

Data dependencies

  • daily_prices

    Adjusted-close OHLCV for every US-listed ticker; primary price feed.

  • transcript_finbert_scores

    Worker data table — see services/worker schema.

Expected edge

Paper window
2002-2008

Analyst Q&A tone predicts 60-day drift independent of the call's stated news content; cleanest signal on calls with multiple analysts asking sharp follow-ups.

Illustrative pattern only

NOT a backtest

Illustrative pattern only — see /app for live backtests and the actual current equity curve.

Related families

transcript finbert qa dispersionText-NLP

Huang-Wang-Yang (2023 CAR) "FinBERT: A Large Language Model for Extracting Information from Financial Text." FinBERT-tone outperforms Loughran-McDonald wordcount on every post-call drift measure tested (3-day CAR R² roughly doubles). Follow-up Chen et al. (2024 JFE) reports ~5% annualised L/S on the Q&A-dispersion sub-signal. Own-history z of per-call Q&A sentiment dispersion (stdev of per-sentence pos-neg score): short when z ≥ +1 (uncertainty), long when z ≤ −1 (confident answers).

transcript qa combativenessText-NLP

Hollander-Pronk-Roelofsen 2010 + Matsumoto-Pronk-Roelofsen 2011: longer analyst questions + lower management answer/question word-count ratios signal skepticism and predict near-term underperformance. Two per-call features: avg_analyst_q_len (tokens per analyst turn) and mgmt_response_ratio (CEO/CFO tokens / analyst tokens). z-score within own history; SHORT when current call's combativeness z > 1.0 (84th percentile). Long leg disabled because positive responses are noisy (good news makes execs talkative).

transcript cfo qa defensivenessText-NLP

When a CFO oscillates between positive boilerplate and negative-tone clarifications during analyst Q&A (high cfo_qa_dispersion in FinBERT scores), they are typically defending weak fundamentals — predicting negative 60-90 day drift independent of the miss/beat itself. The signal isolates *defensive non-answers*: sentence-level CFO tone variance during the Q&A section, z-scored against the firm's own call history.

Explore transcript analyst frustration on alphactor.ai

See which tickers this family is currently firing on, with live signals and rankings.

For informational and educational purposes only. Not financial advice. Learn more