Taiwan Event-Study Stats — 20+ event types, historical base rate
After a type of event, how did prices historically move on average? The event-study card computes historical abnormal-return base rates for 20+ event types, with CAR / BHAR dual benchmarks, sample sizes and intervals.
What events are covered
Institutional streaks, insider filings, block trades, pledge changes, dividend policy, buybacks, revenue surprises — 20+ types, each with average abnormal return and hit rate across holding windows.
Why both CAR and BHAR are shown
Cumulative abnormal return and buy-and-hold abnormal return measure the same event under different compounding assumptions, and they can disagree. Publishing only the more favourable one is how event studies become marketing.
We list both, with sample sizes and windows, and flag when the two benchmarks point in opposite directions. These are historical base rates, not forecasts: a positive average tells you what happened after past instances, not what will happen after the next one.