Free lesson · Alternative data
An event return needs a predeclared benchmark
Open interactive lessonPractice calculationsExplore labs
Start with the idea
An asset moving after a news event does not mean the entire move was caused by that event.
Symbols, units & horizon
- r_i: asset simple return over a fixed event window
- r_m: benchmark return over exactly the same window
- α: pre-event fitted expected intercept return for that window
- β: dimensionless market exposure fitted before the event
- AR: abnormal-return fraction
When and why to use this
Assess whether a news, filing or alternative-data release coincides with returns beyond a simple benchmark.
An asset moving after a news event does not mean the entire move was caused by that event.
Compare the asset return with a benchmark fitted before the event. A market model subtracts its intercept and market exposure over the same window. Overlapping news, trading halts and asynchronous prices can contaminate the comparison.
A descriptive event study is not a causal identification strategy. Trading also requires that the news and feature were available before the measured entry price, including processing and order delay. A delayed market reaction should be tested against alternatives rather than presumed.
An event return needs a predeclared benchmark
- Estimate benchmark parameters on data preceding the event and freeze them.
- Compute the benchmark-predicted return α+βr_m over the event window.
- Subtract it from the asset return; interpret the residual descriptively unless an identification design supports causality.
Asset return .03, benchmark .01, α=0 and β=1.2 give expected .012 and abnormal return .018, or 1.8%.
Apply it in a strategy
- Freeze inputs at the stated decision time and record their units.
- Assess whether a news, filing or alternative-data release coincides with returns beyond a simple benchmark.
- Recompute the example, then change the material assumption and explain the difference.
Research deliverable
An event return needs a predeclared benchmark: produce the worked calculation, a timestamped input record and a written decision addressing this limitation: Confounding events and benchmark misspecification prevent the residual from being interpreted automatically as a causal effect.
These are synthetic mechanics examples, not historical performance or paper replications. Module evidence and research boundaries record the 12 September 2026 review.
Python implementation
Self-contained teaching example. Python 3.10+; dependencies and input conventions are shown in the code and notation. Run in your own Python environment.
# Python 3.10+; standard library and NumPy only.
# Synthetic teaching inputs; conventions and units are defined in the notation above.
def abnormal_return(asset_return,market_return,alpha,beta):
return asset_return-(alpha+beta*market_return)
assert abs(abnormal_return(.03,.01,0,1.2)-.018)<1e-12
print(abnormal_return(.03,.01,0,1.2))Continue learning
Alternative Data: Measurement, Text & Incremental Value — all lessons- From a sampled panel to a population estimate
- A reproducible dictionary score for text
- An event return needs a predeclared benchmark
- Availability delays and signal decay
- Noisy proxies and attenuation
- Measure improvement against a frozen baseline
- From forecast accuracy to a costed decision
- A data investment includes coverage, access and ongoing costs
Quantitative finance and development glossary · Python resources and libraries · Research sources and limitations