RegulationRules, enforcement & market structure
Study links major album release days to 15% rise in US traffic deaths
Research by Vishal Patel, Anupam Jena and colleagues found that US traffic deaths averaged 139 on release days of the 10 most-streamed albums of 2017–2022, compared with 121 on surrounding days. The authors see this as a sign of smartphone distraction.
From the Klement on Investing article of November 20, 2026 and cited sources: The study links NHTSA's census of fatal crashes (233,809 deaths from 2017 to 2022) with daily Spotify data. [3] [6] [2]
SOURCES[3] NBER w34866: Smartphones, Online Music Streaming, and… nber.org[6] HealthDay: Music streaming linked to rise in traffic deaths healthday.com[2] KEYT/CNN: The day big albums start streaming, traffic… keyt.com
AI-assisted, not human-edited. Methodology Corrections Who owns this site
- What happened
- The study links NHTSA's census of fatal crashes (233,809 deaths from 2017 to 2022) with daily Spotify data. [3] [6] [2]
- Original source
- Klement on Investing article, independent perspective
- Brief by
- AT Pulse research desk AI-assisted; publisher and contact on the desk page
- When
- November 20, 2026 article published
- Covered
- Nov 20 edition
- Evidence
- 7 sources across 7 sites
- 1Klement on Investing: Taylor Swift kills peopleklementoninvesting.substack.com
- 2KEYT/CNN: The day big albums start streaming, traffic…keyt.com
- 3NBER w34866: Smartphones, Online Music Streaming, and…nber.org
- 4Harvard Gazette: You heard it first, and then diednews.harvard.edu
- 5PhillyVoice: Traffic deaths and new music releasesphillyvoice.com
- 6HealthDay: Music streaming linked to rise in traffic deathshealthday.com
StrongHow it’s measured
Substack · 14 likes · 3 comments
Reach blends public engagement (views, likes, comments, shares, stars) from the platforms where the story or its original source lives: YouTube, Hacker News, GitHub, Substack and similar. A video and the article written from its transcript share one score, taken from the video. Platforms with no data are left out rather than counted as zero, and the age of a story never lowers its score.

The study links NHTSA's census of fatal crashes (233,809 deaths from 2017 to 2022) with daily Spotify data. On major album release days, streaming rose by roughly 40% and traffic fatalities by about 15%, or around 18 extra deaths per release. The findings appeared as an NBER working paper in February 2026 and in JAMA Network Open in August. [3] [6] [2]
The design is an event study that compares each release day with the 10 days before and after it. The effect was larger among younger drivers, solo drivers and sober drivers. It was not concentrated at night, which argues against late-night partying as the explanation. [2] [3]
The researchers never observed phone use in any individual crash. The link from release to phone use to crash is inferred. Jena said the mechanism could be volume, novelty or handling the phone, and the authors present the result as an association. [2] [4]
Klement's satirical post jokes about banning driving on release days and rounds the effect up to about 20 deaths. Its humour makes the point that observational data cannot show that phones caused the crashes. [1]
a peer-reviewed association between release-day streaming spikes and fatal crashes. Implication: driving-mode defaults and low-interaction interfaces matter for streaming and in-car software.
Read the full assessment
The authors note that future in-car AI could either reduce or increase distraction.
Executive brief
The joke headline sits on a real study. Klement's satirical post, "Taylor Swift kills people", draws on research by Vishal Patel, Anupam Jena and colleagues at Harvard. They found that US traffic deaths rose about 15% on the release days of the 10 most-streamed albums of 2017–2022: an average of 139 deaths versus 121 on surrounding days (KEYT/CNN). The authors say this points to smartphone distraction beyond calls and texts. The post's own satire makes the needed caveat: the data is observational and cannot prove that phones caused the crashes.
What changed and event timeline
Working paper released
NBER published "Smartphones, Online Music Streaming, and Traffic Fatalities" by Patel, Worsham, Liu and Jena. It reports streaming up nearly 40% and fatalities up nearly 15% on major release dates ().
Harvard explains the mechanism
The Harvard Gazette reported 182 excess deaths across 10 releases. Patel argued that many people doing the same distracting thing at once adds up to large risk on the road ().
Wider coverage
PhillyVoice named Midnights, Certified Lover Boy and Un Verano Sin Ti among the albums studied. It also reported that Patel began the project after drifting out of his lane while opening Spotify ().
Peer-reviewed version
JAMA Network Open published the study. CNN coverage reported subgroup results and an outside researcher's concerns about uncontrolled holidays ().
Satire
Klement's post jokes about bans on driving on release days, rounding the effect to "about 20" extra deaths ().
Capabilities and access
Read the full section
- Data: NHTSA's Fatality Analysis Reporting System (FARS), a census of fatal US crashes, linked to daily Spotify streaming data. The analysis covers 233,809 fatalities from 2017 to 2022 (HealthDay).
- Access: the working paper is listed at NBER and RePEc. The final version is in JAMA Network Open.
- No AI model or software product is involved.
Technical analysis for researchers and developers
- Each release day is compared with the 10 days before and after.
- Dose proxy: release-day streaming, up 43% (Harvard Gazette).
- Heterogeneity: larger effects among younger drivers, solo drivers and sober drivers. No nighttime concentration, which argues against a late-night-party explanation (KEYT/CNN).
Read the full section
- Design: an event study. Each release day is compared with the 10 days before and after. The arbitrary timing of releases serves as an outside shock to phone use (NBER).
- Dose proxy: release-day streaming, up 43% (Harvard Gazette).
- Heterogeneity: larger effects among younger drivers, solo drivers and sober drivers. No nighttime concentration, which argues against a late-night-party explanation (KEYT/CNN).
- Gap: phone use is not observed at the crash level. The causal chain from release to phone use to crash is inferred.
Claims and evidence
No independent replication was found in the reviewed sources.
Read the full section
| Claim | Evidence |
| Release-day deaths rise about 15% (139.1 vs 120.9) | Peer-reviewed; HealthDay |
| "About 20" extra deaths per release | Rounded by the satirist; the study reports 18.2 (Klement) |
| Smartphones are the cause | Authors' interpretation. Jena said the mechanism is unknown: volume, novelty or handling the phone (KEYT/CNN) |
No independent replication was found in the reviewed sources.
Context and prior work
- Around 35,000 Americans die in traffic each year. More than 3,300 deaths in 2022 involved distracted driving (PhillyVoice).
- Jena's group often uses "natural experiment" date designs. The study's novelty is its use of everyday media behaviour to measure phone use while driving, which researchers cannot test experimentally for ethical reasons (Harvard Gazette).
Read the full section
- Around 35,000 Americans die in traffic each year. More than 3,300 deaths in 2022 involved distracted driving (PhillyVoice).
- Jena's group often uses "natural experiment" date designs. The study's novelty is its use of everyday media behaviour to measure phone use while driving, which researchers cannot test experimentally for ethical reasons (Harvard Gazette).
Limitations, safety and contested findings
- The study does not confirm phone use in any crash and excludes nonfatal crashes. Spotify covers only about 29% of US streaming users (KEYT/CNN).
- IIHS researcher Ian Reagan noted that non-federal holidays such as Halloween may not be controlled for (KEYT/CNN).
- The sample is only 10 events, and the data ends in 2022, before newer vehicle safety technology (Harvard Gazette).
Read the full section
- The study does not confirm phone use in any crash and excludes nonfatal crashes. Spotify covers only about 29% of US streaming users (KEYT/CNN).
- IIHS researcher Ian Reagan noted that non-federal holidays such as Halloween may not be controlled for (KEYT/CNN).
- The sample is only 10 events, and the data ends in 2022, before newer vehicle safety technology (Harvard Gazette).
- The satire mocks a causal reading, but the authors themselves present the result as an association.
Business and practitioner implications
- Streaming and in-car software teams: the study links release-day engagement spikes to measurable road harm.
- Automakers and voice/AI assistant builders: the authors say future AI in cars could either reduce or increase distraction (Harvard Gazette).
- Insurers and safety analysts: release calendars are a predictable, if small-sample, risk marker.
Read the full section
- Streaming and in-car software teams: the study links release-day engagement spikes to measurable road harm. Driving-mode defaults and low-interaction interfaces are design choices, not decorations.
- Automakers and voice/AI assistant builders: the authors say future AI in cars could either reduce or increase distraction (Harvard Gazette).
- Insurers and safety analysts: release calendars are a predictable, if small-sample, risk marker.
- No market price data or company filings bear on this story.