Technology Layoffs Event Study

A collaborative event study of short-window market reactions to technology-sector layoff announcements, and an exercise in separating descriptive evidence from causal claims. Market-model abnormal returns over a three-day window are tested with one-sample and subgroup means, multivariate OLS with robust inference, logistic classification, and a 10,000-draw placebo simulation. The average reaction is modest and observable characteristics explain little of the cross-sectional variation. I am one of five credited contributors; the public project record assigns data provision and cleaning to another contributor and identifies the remaining work at group level, and this entry preserves that boundary rather than overstating ownership.

Question

Do publicly traded technology companies experience a systematic three-day abnormal return around layoff announcements, and do observable announcement characteristics explain its direction or magnitude?

My role

Academic group project, IE University; one of five contributors. Credit for the analysis and interpretation is shared at group level; data provision and cleaning were another contributor’s workstream.

Method

Python · Econometrics · OLS · Simulation

Tools

Python in Jupyter notebooks · pandas, NumPy, SciPy, statsmodels, scikit-learn · yfinance for market prices · matplotlib and seaborn

Evidence

Technology-sector layoff announcements matched to listed tickers, with market prices retrieved at run time; three-day event window.

Result

−0.5%Mean three-day abnormal return

0.52

In-sample AUC

10,000

Placebo draws

Findings

• Mean three-day abnormal return of approximately −0.5%
• In-sample classification AUC of approximately 0.52, not evaluated on a holdout
• Differences between smaller and deeper cuts are descriptive and statistically nonsignificant
• A 10,000-draw placebo simulation used as an interpretation check, not as support

Output

Two reproducible notebooks, an executive summary and a presentation; two-page dossier (PDF).

BoundaryDescriptive market evidence—not a claim that layoffs caused observed returns or that the classifier predicts out of sample.