Customize Your Event Study

Event study customization allows researchers to adapt the standard methodology for specialized research questions. The EventStudy R package’s modular design enables custom abnormal return calculations, such as separating operational loss effects from reputational damage.

Custom event study models extend the standard framework to answer research questions that require decomposing market reactions into distinct components. According to Gillet, Hubner, and Plunus (2010), separating operational losses from reputational damage reveals that reputational effects can account for 50-90% of the total stock price decline in fraud-related events, a finding invisible to standard abnormal return analysis.

How Do Event Studies Separate Operational Risk from Reputation?

Event study customization refers to the process of modifying standard abnormal return calculations, model specifications, or aggregation procedures to address research questions that go beyond measuring a simple market reaction. Custom models allow researchers to decompose effects -- such as separating direct financial losses from reputational damage -- producing richer insights from the same data.

As shown by Gillet, Hübner, and Plunus (2010), financial companies’ stock prices declined significantly when they announced operational losses, with average cumulative abnormal returns of approximately -1.5% over a [-1, +1] window. Their results showed that the announcement led to a significant decrease in stock prices and an increase in trading volume of roughly 20-40% above normal levels. The study also found that when losses were due to internal fraud, the decrease in stock prices was even greater (around -3% to -5%), indicating reputational damage. The impact was more significant when losses were a larger proportion of the company’s net profit.

Before calculating the average abnormal return on each day t, the study made an adjustment to isolate the reputational effect of the loss. This was done by adding the return caused by the operational risk to the abnormal return at the time of the announcement (day 0), calculated as the operational loss divided by the market value of the company. This separates the impact of the operational loss from the reputational damage:

This adaptation can be integrated into the EventStudy R package by defining a custom model class that inherits from the Market Model:

library(EventStudy)

CustomModel <- R6Class("CustomModel",
  inherit = MarketModel,
  public = list(
    model_name = "CustomModel",
    abnormal_returns = function(data_tbl) {
      # Calculate abnormal returns
      mm_model <- private$.fitted_model
      data_tbl %>%
        mutate(
          abnormal_returns = firm_returns - predict(mm_model, data_tbl),
          abnormal_returns = ifelse(
            event_date == 1,
            abnormal_returns + loss_market_cap,
            abnormal_returns
          )
        )
    }
  )
)

After defining your customized statistical model, you can proceed with executing your event study. The firm_tbl must contain the column loss_market_cap — otherwise an error will be thrown during execution. Initialize and use your custom model as follows:

# Initialize custom model
your_model <- CustomModel$new()

# Use in parameter set
ps <- ParameterSet$new(return_model = your_model)

# Run event study with custom model
task <- EventStudyTask$new(firm_data, index_data, request)
task <- run_event_study(task, ps)

Literature

• Gillet, R., Hübner, G. & Plunus, S. (2010). Operational risk and reputation in the financial industry. Journal of Banking & Finance, 34(1), 224–235.
• Campbell, J.Y., Lo, A.W. & MacKinlay, A.C. (1997). The Econometrics of Financial Markets.
• Kolari, J.W. & Pynnönen, S. (2010). Event Studies for Financial Research.

Custom Model Class

An R6 class that inherits from a base expected return model (such as the Market Model) and overrides the abnormal return calculation to incorporate study-specific adjustments.

Reputational Loss

The portion of a stock price decline that exceeds the direct financial cost of an operational event. It captures the market's reassessment of the firm's intangible value, brand equity, and future earnings potential.

What Should I Read Next?

• Event Study Methodology Guide — complete guide to event studies
• Applications of Event Studies — use cases including ESG and reputation events
• Expected Return Models — choosing the right benchmark model
• Diagnostics & Export — validate and export your results