Skip to content
All library documents

Pooling Event-Study Returns Across Firms with Different Event Dates

Article Quant Q&A · Author: amber
Read the original

Summary

The document addresses event studies where many companies experience events on different calendar dates. Its central method is to align each event as time zero, estimate the normal-return model in a pre-event window, and express returns on surrounding days relative to each company’s event date. Abnormal returns can then be collected across events for a chosen relative day and tested against zero or another benchmark.

The answer sketches a pre-event estimation window and suggests a t-test on the pooled abnormal returns. It also points to an R event-study package as a way to process multiple tickers, though the respondent had not established that it scales to the full sample described. The guidance is brief: it does not discuss dependence among repeated events for the same firm, cross-sectional correlation, overlapping event windows, model selection, or adjustments to inference. Those issues can affect whether a simple pooled t-test is appropriate.

Key ideas

  • Align each observation by its own event date and treat that date as relative day zero.
  • Estimate expected returns in a pre-event period, then calculate abnormal returns around each event.
  • Pool abnormal returns at a selected relative day and test them against a benchmark.
  • A basic t-test is suggested, but repeated events and cross-sectional dependence may complicate inference.
  • An R package is mentioned, with no evidence about its capacity for the full sample.

Tags

EquitiesEvent-drivenStatisticsBacktesting
Full text
# How to conduct an event study for multiple companies with different event dates?


# How to conduct an event study for multiple companies with different event dates?












I have a sample of 300 companies over the period of 5 years. Each company has one event per year and all event dates are almost different. Is there any short way to do event study instead of doing for each company and each year in Excel?

## Answer by JohnDoe (score 2)

https://quant.stackexchange.com/a/44728

You just have to treat every event as day zero and look at the other days relative to the event day. So the estimation period will be something like days -260 to -11. After that you can calculate tej abnormal returns around the event date.

Then, just take for example every day 1 abnormal return (5 tines 300 in your example) and test them against zero or whatever you want to check. The simplest thing would be a t-test with 1500 abnormal returns.

## Answer by Con Fluentsy (score 0)

https://quant.stackexchange.com/a/68090

If you are familiar with R statistical language a full library of event-study packages is implemented, the one I have used and it works well and it works with multiple stock tickers, I am not sure if there is a limit as 10 or so is enough for me, and I have not extended to 300, as you wish too. But it is simple quick efficient, the library is from the R cran repository called Estudy2.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.