Home / Financial Tools / Event Study Abnormal Return & Market Model Significance Tester

Graduate-Level Modeling · Capital Markets Research

Event Study Abnormal Return & Market Model Significance Tester

The market moved. The stock moved more. This tool tells you exactly how much more, and whether that gap is a real signal or just noise.

How To Use This Model

Reading This Tool

Enter the market model parameters estimated over a clean prior period, then the stock and market's actual returns across a five-day event window.

The tool computes each day's abnormal return against the market model's expectation, builds the cumulative abnormal return, and tests whether it's statistically distinguishable from zero, the standard event-study methodology behind most empirical corporate finance and accounting research.

Market Model (Estimated Over Prior Period)

Alpha (Daily, bps) 2 bps

Beta 1.20

Residual Std. Dev. (σ_ε, Daily) 1.20%

Event Window Returns (%)

Day	Stock Return	Market Return
t−2
t−1
t=0
t+1
t+2

Day 0 is the event date, an earnings announcement, M&A disclosure, or any other information event you're studying. Returns are expressed in percent.

Event Study Result

Cumulative Abnormal Return (CAR)

t-Statistic

Significant At 5%?

Significant At 1%?

Cumulative Abnormal Return Build-Up

Empirical Capital Markets Research

The Core Formulas

Expected Return_t = α + β·Market Return_t
Abnormal Return_t = Actual Return_t − Expected Return_t
CAR = Σ Abnormal Return_t Over The Event Window
t = CAR / (σ_ε × √N)

When To Actually Use This Method

Teaching empirical event-study methodology in a capital markets, corporate finance, or accounting research course.
Quantifying the market's reaction to a specific disclosure, earnings surprise, M&A announcement, guidance change, regulatory action, in a research or investor-relations context.
Sanity-checking a published event-study result's CAR and significance calculation before citing it further.

Key Assumptions & Limitations

Assumes market model residuals are independently and identically distributed, ignoring cross-sectional correlation, real large-sample event studies aggregate across many firms and adjust for this.
A single-firm event study like this one is illustrative, published event studies pool dozens or hundreds of comparable events to get statistical power a single observation can't provide.
Alpha and beta are taken as given here, in practice they're estimated by regressing the stock's returns on market returns over a 100-250 day estimation window ending well before the event window begins.

Foundational Reference

MacKinlay, A. C. (1997). Event Studies in Economics and Finance. Journal of Economic Literature, 35(1), 13-39.

Testing whether two series actually move together long-run?

Open The Cointegration Tester →