Home / Event study statistics / Non-routine material information
Non-routine material information
n=1,391as of 2026-07-09
Material information filings left after routine periodic filings are excluded.
| Window | Benchmark | n | Mean abnormal return | Median | q value | Significant |
|---|---|---|---|---|---|---|
| Event day | BHAR | 1,164 | +0.56% | +0.16% | 0.000 | Yes |
| 5 trading days after | BHAR | 1,044 | +1.79% | +0.21% | 0.000 | Yes |
| 20 trading days after | BHAR | 614 | +5.40% | +1.42% | 0.000 | Yes |
Method: the filing's public-availability date is day 0, with a conservative entry on the following trading day. CAR (cumulative abnormal return) is estimated with a market model and calibrated by a placebo test that keeps the event dates but draws random same-day tickers. BHAR (buy-and-hold abnormal return) is measured against size- and liquidity-matched control stocks. All tests carry a multiple-testing correction (BH-FDR). Where the two benchmarks disagree, both are shown as they are — we do not pick the flattering one.
Frequently asked questions
How do Taiwan share prices behave after a non-routine material filing goes out?
Across n=1,391 historical events measured by this site's event study engine, mean abnormal return +0.56% on the event day (BHAR benchmark). Both CAR and BHAR benchmarks are presented, flagged where they disagree. This is not investment advice.
How large is the sample when a non-routine material filing goes out?
1,391 events, as of 2026-07-09, of which 6 statistical cells clear the significance threshold.
Other event categories
The above is a statistical view of historical events (base rates). It is not investment advice and not a trading signal; a historical distribution does not guarantee the future repeats it. Source: This site's event study engine (computed in-house from official public-source event data; editorial copyright)