You are here

Event Study Assumptions

Every event study represents a joint test of the research hypothesis, the particular model of expected returns used, and methodological assumptions (Brown and Warner, 1980). As for the methodological assumptions, the following three are most central:

  1. The stock returns in the event window of a particular event study application accurately reflect the economic impact of the event (capital market efficiency)
  2. The event is unexpected and has not yet been factored into the stock price
  3. There are no other events during the event window, which could be responsible for the stock price change

Depending on the expected return model used, more assumptions need to be met. For the most common model, the market model, for example, the relationship between the stock and the market needs to remain stable throughout the estimation and the event window. Only then, the alpha and beta factors, which were established with a regression analysis during the estimation window, can be used to predict expected returns for the event window.

The methodological assumptions raise several questions the researcher must review:

Question Implication
Is the stock of the analyzed firm frequently traded?  Infrequent trading of the firm's stock may lead to problems in deriving the estimation parameters $\alpha$ and $\beta$ of the market model. Further, infrequent trading suggests that the capital market might not be efficient, questioning the validity of the stock price reaction.
Is the capital market represented by the reference index liquid and shows sufficient trading volume? similar to the above
Are the time series of prices between the stock and the reference matching? Mismatches in the time series of returns in the stock and market returns throughout the estimation window may lead to overall shorter estimation periods and potentially biased parameters.
Has information leakage taken place prior to the event? If information about the event has leaked to capital markets prior to the event window, the CAR of the event is not correct since a certain part or the totality of the event has already been priced into the stock price during the estimation window.
Have there been other events during the event window that could be responsible for the analyzed firm's stock price changes?

For smaller sample studies and event studies with a single firm/event combination, confounding events may void the validity of results. In large sample studies, the adverse effects of confounding events may be sufficiently 'corrected' by creating mean values over large numbers of observations,

Is the chosen reference index the best correlate to the firm's stock price?  If the chosen reference index is not the best possible correlated, analysis results may turn out biased. 
Has the relationship between the reference index and the firm's stock price change over the estimation period? The $\beta$ factor that is calculated from the estimation period would be biased. Predictions of normal returns would turn out incorrect and with them also the abnormal returns.