使用R语言eventstudies包复现eventstudy函数时遇阻求助

Let’s walk through how to get your event study code working. I’ve tackled similar problems with this package before, so here are the key areas to check:

1. Verify Your Data Formats Match Package Requirements #

The eventstudy() function is picky about input structures—let’s confirm your data fits:

• StockPriceReturns (firm.returns): This should be either an xts object or a data frame where:
  • Rows represent dates (must be a standard Date type, not character strings)
  • Columns are individual stocks, with column names matching the firm.identifier values in your SplitDates data
  • All values are numeric (no character-based returns or missing values). Run str(StockPriceReturns) and anyNA(StockPriceReturns) to validate.
• SplitDates (event.list): This needs to be a data frame with at least two columns:
  • firm.identifier: Exact matches to the column names in StockPriceReturns (e.g., stock tickers)
  • event.date: A Date-type column with the date of each stock’s split event
  • Double-check that each event date falls within the date range of your StockPriceReturns data—events outside this range will cause failures.

2. Double-Check Function Parameters #

Let’s break down the parameters you’re using to make sure they align with your data:

• event.window = 7: This sets a total window of 7 days (3 days before the event, the event day, 3 days after). Ensure your StockPriceReturns has data for every day in this window for each stock in SplitDates—missing dates here will break the calculation.
• to.remap = TRUE + remap = "cumsum": This converts single-period returns to cumulative returns. Make sure your raw returns are single-period (not already cumulative)—using pre-cumulated data here will give nonsensical results.
• inference = TRUE + inference.strategy = "bootstrap": Bootstrap inference requires a sufficient sample size (enough stocks/events) and complete data. If you’re working with a small dataset, try running the code first with inference = FALSE to rule out issues with the bootstrap step.

3. Test with the Package’s Built-In Sample Data #

First, confirm the package itself works by running the exact example code with the provided sample data:

library(eventstudies)
data(StockPriceReturns)
data(SplitDates)

# Run the sample event study
es <- eventstudy(firm.returns = StockPriceReturns, event.list = SplitDates, 
                 event.window = 7, type = "None", to.remap = TRUE, 
                 remap = "cumsum", inference = TRUE, 
                 inference.strategy = "bootstrap")

# Check the results
summary(es)

If this runs successfully, the problem is definitely with your custom data. If it fails, try reinstalling the package with dependencies:

install.packages("eventstudies", dependencies = TRUE)

4. Use Error Messages to Narrow Down Issues #

When you run your code with your own data, R will output specific error messages (e.g., "invalid date format", "firm identifier not found"). These messages are your best clue—if you hit a specific error, cross-reference it with the checks above (e.g., a date error means your event dates or return dates aren’t properly formatted as Date types).

内容的提问来源于stack exchange，提问作者Niall Skelton