Limits of Event Studies for Measuring Private-Company Revenue Effects
Summary
The document considers whether event-study methods can measure the effect of management actions, such as advertising campaigns, on a private company’s revenue rather than on a public company’s stock returns. The proposed approach would compare observed revenue with a baseline that accounts for ordinary patterns, potentially using detrending or seasonal adjustment and estimating cumulative abnormal revenue.
The answer highlights substantial data limits. Financial reporting may provide revenue only at low frequency, leaving too few observations to estimate a reliable baseline or isolate the effect of a brief event. Long historical windows can also become unrepresentative as the business and its environment change, while many other factors influence revenue. Higher-frequency cost-accounting or performance-management data could make the analysis more feasible, but such data lack standardization and are rarely disclosed. The note outlines conceptual feasibility and practical barriers rather than a tested design or specific estimation procedure.
Key ideas
- An event study for a private company could compare revenue with an estimated normal baseline.
- Seasonality and other recurring patterns may need to be accounted for when estimating that baseline.
- Low-frequency financial reporting can leave too few observations to identify an event’s revenue effect.
- Long estimation periods risk becoming unrepresentative as a business and its environment change.
- Higher-frequency internal accounting data may help, but they are rarely standardized or publicly available.
Tags
Full text
# Event studies using revenue data vs. measuring abnormal returns # Event studies using revenue data vs. measuring abnormal returns This may be a silly question, but does there exist a methodology for examining the impact of "events" on companies that are not publicly traded? I suppose it would look at abnormal revenues rather than abnormal returns on the stock market. Say, for example, that a private business wants to analyze the impact of an advertising campaign or promotion (or any other endogenous event). The "event" would be the promotion, and they would analyze any change in revenue from their "normal" levels. I suppose they would use historical revenue data that had been cleaned to remove any possible cyclical or seasonal components (detrended to isolate the impact of the "event"). So this wouldn't have anything to do with investor expectations or whether an event has real information content. I've found a paper: "Event study methodology in marketing". What I'm looking for is something similar but assessing revenues rather than stock prices. The overarching idea being to quantify the impact of management decisions when there is no stock market data. I apologize if this is too far outside the realm of quantitative finance. I learned about event studies through investment analysis and financial economics classes. At the moment I'm wondering what sort of data-analysis is possible for smaller businesses. ## Answer by Constantin (score 1) https://quant.stackexchange.com/a/15643 Revenue data for non-public companies are available only at a very low frequency, based on financial reporting requirements. It would be impossible to have a long enough period to estimate the normal return in the first place, let alone detect the effect that an event on one single day will have on the annual revenues. Also, it would be very difficult to obtain a stationary time series from this, given that for a large enough estimation period (100 years?) the business environment will be transformed completely and there is an infinite number of other factors which have influence the company's profitability. It would in theory be possible to use data from performance management systems (cost accounting rather than financial accounting) which may be available to a company at a much higher frequency, which could then be used for a normal return estimation and the determination of cumulative abnormal returns (CARs). However, the implementation of this will be very difficult or impossible because cost accounting information has virtually no standardisation, and is not disclosed by any companies. Whether to use prices or returns is a completely different question.
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.