Last modified:2026-07-19 17:02:27
Maintainer:Not configured
OperationId:equities_symbol_earnings_event_study
MacKinlay market-model event study around the company's quarterly earnings announcements. Announcement dates are sourced from SEC 8-K item 2.02 filings (the report/announcement date carried by the filing); period, year and quarter reflect that report/announcement (calendar) date, NOT a fiscal quarter-end. Per event: OLS market-model fit on the estimation window [-250, -11] vs market_proxy (default SPY), daily abnormal returns over [0, +20], and CAR over [0,+1], [0,+5], [0,+20]. Cross-event significance is a two-sided one-sample Student-t test (df = n-1, exact incomplete-beta - NOT a normal approximation) plus a percentile bootstrap CI. Returns are SIMPLE (close_t/close_{t-1} - 1) computed on stock + proxy close prices aligned on common trading days. Events with no calendar announcement date are excluded (skip_reason=missing_announcement_date). The retrospective filter (drop events whose announcement is >45 days after period) is effectively inactive here because the announcement date is taken directly from the filing, so period equals the announcement and there is no period-vs-announcement lag to screen. Daily only (interval=1d).