When a company announces an earnings beat, a merger, or a CEO resignation, its stock price moves. But how much of that move is truly due to the event — and how much would have happened anyway, just from normal market fluctuations? That’s exactly the question that Cumulative Abnormal Return (CAR) is designed to answer.
CAR is the foundational metric of event study methodology in finance and academic research. It strips away expected market movement to isolate the pure, event-driven portion of a stock’s performance over a defined window of time. Whether you’re a finance student running a research paper, an analyst studying M&A impacts, or an investor evaluating market efficiency, this guide covers everything you need to know.
What Is Cumulative Abnormal Return (CAR)?
Cumulative Abnormal Return (CAR) is the sum of all abnormal returns for a security over a defined event window. An abnormal return, in turn, is the difference between a stock’s actual return and its expected return, the return it would have earned under normal market conditions without the event.
In plain English: CAR measures how much a stock over- or under-performed expectations because of a specific event.
It was formally introduced in the landmark 1969 study by Fama, Fisher, Jensen, and Roll on stock splits, often called the FFJR study and has since become the standard metric for event study analysis in both academic finance and professional investment research.
Positive CAR
Stock outperformed expectations during the event window, the event had a positive market impact.
Negative CAR
Stock underperformed expectations, the event had a negative or disappointing market impact.
CAR Near Zero
Stock performed roughly as expected, the event had little measurable incremental market impact.
Why it matters: CAR isolates the event’s true contribution to stock performance by removing background market noise. Without this adjustment, you cannot tell whether a stock moved because of the event or simply because the broader market moved on that day.
The CAR Formula Explained
CAR is built from two nested calculations. First, compute the Abnormal Return (AR) for each day. Then, sum those values across the event window to get CAR.
Step 1 — Per Day
ARₜ = Rₜ − E(Rₜ)
Actual Return minus Expected Return for each day t in the event window
Step 2 — Sum Across Window
CAR = Σ ARₜ
Sum of all abnormal returns from event start to event end
What Each Variable Means
| Variable | Meaning | How It Is Obtained |
|---|---|---|
| ARₜ | Abnormal return on day t | Calculated: Rₜ − E(Rₜ) |
| Rₜ | Actual stock return on day t | From price data: (P₁ − P₀) / P₀ |
| E(Rₜ) | Expected return on day t | From CAPM, market model, or mean model |
| CAR | Cumulative abnormal return | Sum of all ARₜ across event window |
| Σ | Summation symbol | Add all daily ARₜ values together |
Quick Example: If a stock had abnormal returns of +1.2%, −0.3%, +0.8%, and +0.5% across a four-day event window, the CAR would be +1.2% + (−0.3%) + 0.8% + 0.5% = +2.2%. The event generated 2.2 percentage points of return beyond what the market expected.
Expected Return Models: CAPM, Market-Adjusted, Mean-Adjusted
The most critical decision in a CAR event study is choosing your expected return model. The expected return is the benchmark, how the stock “should have” performed without the event. Three models are widely used, each with different assumptions and use cases.
CAPM (Capital Asset Pricing Model)
E(Rₜ) = Rf + β × (Rm − Rf)
CAPM accounts for both market movement and the stock’s individual risk sensitivity (beta). It uses the estimation window (a pre-event period) to calculate beta, then applies that beta to the market return during each event window day.
Best for: Academic research, finance dissertations, studies where risk-adjusted expectations matter. The most rigorous and most commonly required model in peer-reviewed event studies.
Market-Adjusted Return Model
E(Rₜ) = Rm (Market Return on day t)
The simplest model, it assumes the expected return on any day is simply equal to the market index return on that day. No estimation window or beta calculation required.
Best for: Quick analyses, preliminary screening, studies where you assume beta ≈ 1, or when estimation window data is limited.
Mean-Adjusted Return Model
E(Rₜ) = Mean(Stock Return during estimation window)
Uses the stock’s own average historical return (from the estimation window) as the expected return. Ignores market movements entirely.
Best for: Studies where market correlation is weak, or when studying idiosyncratic events with no broader market linkage. Less common in published research.
Which model should you use? For academic papers and finance research, CAPM is the gold standard. For quick exploratory analysis or when you lack estimation window data, market-adjusted is the pragmatic choice. All three models are available in the CAR calculator at calcxi.com.
What Is an Event Window?
An event window is the period of trading days, centered around an event date, over which CAR is measured. It is typically expressed as [t₁, t₂] where t = 0 is the event date itself, negative values are days before the event, and positive values are days after.
Example: Event Window [−5, +5]
ESTIMATION WINDOW
e.g. −120 to −21
Used to calculate Beta / Mean Return
PRE-EVENT
t = −5 to −1
Anticipation / leakage
EVENT DATE
t = 0
Announcement day
POST-EVENT
t = +1 to +5
Delayed reaction
Common Event Windows and When to Use Them
| Event Window | Days Covered | Best Used For |
|---|---|---|
| [0, 0] | 1 day | Pure event-day reaction (earnings day, announcement day) |
| [−1, +1] | 3 days | Short-term reaction including day before and day after |
| [−3, +3] | 7 days | Moderate reaction window; most commonly cited in literature |
| [−5, +5] | 11 days | Capturing pre-event information leakage and delayed reactions |
| [−10, +10] | 21 days | Long-term or slow-diffusing events (M&A, regulatory changes) |
Important: The event window and the estimation window must not overlap. The estimation window is used to model normal behavior, so it must represent a “clean” pre-event period without contamination from the event itself. A gap of at least 10–20 trading days between the two windows is standard practice.
How to Calculate CAR Step by Step
Here is the complete manual process for calculating CAR using the CAPM model, the most rigorous approach.
Gather your price data
Collect daily closing prices for both the stock and the market index (e.g., S&P 500) for the full period: estimation window + event window. Include the daily risk-free rate if using CAPM.
Calculate daily returns
Convert prices to returns for each day:
Rₜ = (Price₍ₜ₎ − Price₍ₜ₋₁₎) / Price₍ₜ₋₁₎
Estimate beta from the estimation window
Run an OLS regression of stock excess returns on market excess returns over the estimation window:
(Rₜ − Rf) = α + β × (Rm − Rf) + εₜ
This gives you estimated beta (β) and alpha (α) for the stock.
Calculate expected returns for each event window day
Apply the CAPM formula to each day in the event window:
E(Rₜ) = Rf + β × (Rmₜ − Rf)
Compute abnormal returns
For each day in the event window:
ARₜ = Rₜ − E(Rₜ)
Sum the abnormal returns to get CAR
Add all daily ARₜ values from event start to event end:
CAR = AR₋₅ + AR₋₄ + … + AR₀ + … + AR₊₅
Skip the manual work: All six steps above are automated by the CAR Calculator at calcxi.com. Just upload your CSV with date, stock, and market data, choose your model and windows, and get the full event window table, beta, alpha, and downloadable chart in seconds.
Calculate CAR Instantly — Free Online Tool
Doing a CAR calculation manually in Excel is tedious, error-prone, and time-consuming, especially when you’re managing OLS regression, estimation windows, and multi-day event tables simultaneously. The Cumulative Abnormal Return Calculator at CalcXi handles all of it automatically.
Free CAR Tool · No SignupHow to Use the CAR Calculator (Quick Start)
Upload your CSV
Use the Download Input CSV button for a ready-made template: date, stock, market, rf columns.
Select data type
Choose “Prices” if your data is closing prices, or “Returns” if already converted to decimal daily returns.
Choose model
Select CAPM, market-adjusted, or mean-adjusted depending on your research requirements.
Set your windows
Enter the event date, estimation window start/end, and event window start/end dates.
Click Calculate CAR
Get instant results: final CAR, beta, alpha, full event window table, and running CAR chart.
Download outputs
Export the full results CSV for your research paper or download the CAR chart as a PNG for presentations.
How to Interpret CAR Results
A raw CAR figure means little without context. Here’s how to read it correctly:
Positive CAR (e.g., +3.5%)
The stock returned 3.5 percentage points more than the market model predicted. The event generated positive, statistically meaningful value. Typical in strong earnings beats, positive M&A reactions, or favorable regulatory rulings.
Negative CAR (e.g., −2.1%)
The stock returned 2.1 percentage points less than expected. The event had a negative market impact. Common in earnings misses, CEO scandals, adverse legal outcomes, or macro shocks.
CAR Near Zero (e.g., −0.2%)
The event had no material impact beyond normal market movement. The information may have been already priced in, or the event was deemed immaterial by investors.
Statistical Significance of CAR
A CAR of +3% sounds meaningful, but is it statistically significant or just random variation? In academic event studies, CAR is typically tested using a t-test or z-test. A p-value below 0.05 (95% confidence) is the standard threshold for declaring a result significant.
Practical Rule of Thumb: For a single-stock event study, a CAR exceeding ±2–3% is typically considered economically meaningful. For cross-sectional studies averaging CAR across many events, even smaller values (±0.5–1%) can be statistically significant with sufficient sample size.
When and Why CAR Is Used in Event Studies
CAR is the core output metric of event study methodology, a research design that measures the financial market’s reaction to a specific, identifiable event. It is used across finance, economics, law, and accounting research. Here are the most common real-world applications:
Mergers & Acquisitions
Measure value creation (or destruction) for acquiring and target company shareholders at M&A announcement. A cornerstone of M&A finance research.
Earnings Announcements
Quantify the market’s reaction to earnings beats and misses, separating the expected component from the surprise component of earnings releases.
CEO & Leadership Changes
Assess whether markets react positively or negatively to executive appointments, resignations, or sudden departures and whether the reaction differs by industry.
Regulatory & Legal Events
Study how FDA approvals, antitrust rulings, court decisions, or new regulations affect firm value. Widely used in law and economics research.
Dividend Announcements
Evaluate market reactions to dividend initiations, changes, omissions, or special dividends, testing signaling theory and market efficiency hypotheses.
Product Launches & IPOs
Measure market enthusiasm for new products, service expansions, or initial public offerings, particularly in technology and pharmaceutical sectors.
Central Bank Announcements
Analyze how interest rate decisions, QE announcements, or forward guidance affect specific sectors or individual stocks beyond broad market moves.
ESG & Sustainability Events
A growing area: studying market reactions to ESG disclosures, environmental incidents, or sustainability commitments, testing whether ESG events move stock prices.
CAR vs Abnormal Return: Key Differences
These two terms are often confused, but they measure different things:
| Metric | What It Measures | Time Scope | Best Used When |
|---|---|---|---|
| Abnormal Return (AR) | Unexpected return on a single day | One day / one period | Studying a single-day market reaction (e.g., the exact announcement day) |
| Cumulative Abnormal Return (CAR) | Total unexpected return across multiple days | Multi-day event window | Capturing the full market impact of an event including pre- and post-event drift |
In practice, most event studies report both the daily ARs to show the day-by-day pattern of the market reaction, and the final CAR to summarize the net impact across the entire window. The CalcXi CAR calculator provides both in its event window table.
Limitations of CAR
Like any financial metric, CAR has important limitations that researchers and analysts should be aware of:
Model Dependency
CAR results can differ significantly depending on which expected return model you use. CAPM assumes a linear market relationship which may not hold in all conditions.
Confounding Events
If other significant events (macro news, industry shocks) occur during the event window, separating their impact from the target event becomes difficult or impossible.
Information Leakage
Markets often react before the official announcement date due to leaks or informed trading. The “true” event date may differ from the public announcement date.
Thin Trading & Illiquidity
For small-cap or illiquid stocks, price movements may be driven by trading volume rather than information, making CAR unreliable.
Short-Window Bias
Very short event windows may miss delayed market reactions. Very long windows risk contamination from unrelated subsequent events.
CAPM Limitations
If using CAPM, all the assumptions of CAPM apply, including constant beta and a single-factor market model. Multi-factor models (Fama-French) can provide more robust benchmarks.
Worked Example: Earnings Announcement Event Study
Let’s walk through a simplified example. Suppose a company (Stock XYZ) announces earnings on Day 0. We want to measure the event’s impact using a [−3, +3] event window and the market-adjusted model. Beta has been estimated at 1.15 from the estimation window.
| Day (t) | Stock Return (Rₜ) | Market Return (Rm) | Expected E(Rₜ) | Abnormal ARₜ | Running CAR |
|---|---|---|---|---|---|
| −3 | +0.4% | +0.3% | +0.35% | +0.05% | +0.05% |
| −2 | +0.8% | +0.2% | +0.23% | +0.57% | +0.62% |
| −1 | +1.5% | +0.1% | +0.12% | +1.38% | +2.00% |
| 0 ★ EVENT | +4.2% | +0.4% | +0.46% | +3.74% | +5.74% |
| +1 | +0.6% | +0.5% | +0.58% | +0.02% | +5.76% |
| +2 | −0.2% | −0.3% | −0.35% | +0.15% | +5.91% |
| +3 | +0.1% | +0.2% | +0.23% | −0.13% | +5.78% |
| Final CAR [−3, +3] | +5.78% | ||||
Interpretation: The earnings announcement generated a CAR of +5.78% over the seven-day event window. The bulk of the abnormal return (+3.74%) arrived on the event day itself, with some pre-announcement drift visible on days −2 and −1, possibly suggesting information leakage. The market quickly absorbed the information, with near-zero abnormal returns on days +1 through +3.
Run Your Own CAR Event Study in Minutes
Upload your price data CSV, set your event date and window, and get a complete event study output, including beta, alpha, running CAR table, and downloadable chart. Free, no login required.
Frequently Asked Questions
The Bottom Line
Cumulative Abnormal Return is one of the most powerful and widely-used metrics in empirical finance. By stripping out expected market movement, CAR isolates the true, event-driven impact on stock performance, making it indispensable for event study research in M&A, earnings, regulation, corporate governance, and beyond.
The manual calculation, gathering data, running OLS regression, building the event window table day by day, is both time-consuming and error-prone. That’s exactly what the free CAR calculator at CalcXi eliminates.
Whether you’re completing a finance dissertation, conducting professional research, or simply exploring how a specific corporate event affected stock performance, the CalcXi CAR calculator gives you a complete event study output — CAPM model, beta, alpha, running CAR table, and downloadable chart, in minutes.
Published on calcxi.com · Updated June 2026 · Topic: Quantitative Finance · Event Studies