Graduation Year

2026

Date of Submission

4-2026

Document Type

Open Access Senior Thesis

Degree Name

Bachelor of Arts

Department

Economics

Reader 1

Darren Filson

Abstract

Opening weekend box office results are among the most visible and immediately quantifiable signals of film performance. This thesis examines whether opening weekend box office surprises predict movie studios’ cumulative abnormal stock returns (CARs) and whether post-announcement drift exists following these events.

I construct a dataset of 52 wide-release films from 2025 released by four major publicly traded studios: Disney (DIS), Warner Bros. Discovery (WBD), Universal/Comcast (CMCSA), and Sony (SONY). Box office surprise is measured as the difference between actual and analyst-forecasted opening weekend revenue — where expectations are drawn from Boxoffice Pro, an industry forecasting service that publishes pre-release consensus estimates — scaled by expected revenue, with alternative specifications based on dollar differences and surprise scaled by production budget. Using WRDS Eventus, I estimate abnormal returns with the Fama-French three-factor model and compute cumulative abnormal returns (CARs) over multiple event windows spanning day −10 to day +30 (where Day 0 is the opening Friday). The primary measure of immediate market reaction is CAR(−1,1), while post-event drift is analyzed over horizons up to 30 trading days.

Opening weekend outcomes are associated with economically meaningful but highly heterogeneous stock price reactions. The mean CAR(−1,1) is −1.19%, but the distribution is sharply bifurcated: events above the median generate average returns of +1.54%, while those below generate −3.92%. Despite this dispersion, box office surprise does not reliably predict announcement-window returns. Percentage-based surprise measures are statistically insignificant, and a sign-based test shows no relationship between the direction of surprise and the direction of stock returns. Notably, surprise measured in dollar terms is a significant negative predictor of announcement-window returns (β = −0.002, p < 0.05), a counterintuitive result suggesting either that Boxoffice Pro forecasts are flawed proxies for investor expectations, or that large positive box office outcomes may not materially shift valuations for high-budget releases where investor expectations were already elevated. In bivariate specifications, positive box office surprises are associated with higher CAR(2,25), CAR(3,25), and CAR(3,30); however, this relationship disappears when controlling for film characteristics, and appears to be driven by a small number of large positive surprise events, suggesting idiosyncratic outliers rather than systematic mispricing. Overall, opening weekend performance does not provide a consistent or robust signal for predicting abnormal returns.

Download

Share

COinS