How Dangerous Are Drinking Drivers?
DOI: 10.1086/323281
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the difficulty of measuring the specific risks posed by drinking drivers and the resulting externalities, a problem complicated by the lack of data on the total fraction of drivers on the road who have consumed alcohol. Previous research relied on costly and infrequent random roadblocks, which suffered from sample selection bias due to driver refusal rates and limited generalizability. To overcome these limitations, the authors develop a methodology that relies exclusively on readily available data from fatal crashes, specifically leveraging the statistical properties of two-car collisions to identify both the prevalence of drinking drivers and their relative crash risk. The authors construct a model based on the binomial distribution of driver interactions, assuming that drivers mix randomly on the road and that fatal crashes result from a single driver’s error. By analyzing the composition of two-car fatal crashes—specifically the frequencies of crashes involving two drinking drivers, two sober drivers, or one of each—the authors can separately identify the ratio of drinking to sober drivers on the road and the relative likelihood of a drinking driver causing a fatal crash compared to a sober driver. This identification strategy exploits the nonlinearities inherent in the binomial distribution, allowing for estimation without arbitrary functional forms or external data on driver populations. The model is applied to U.S. fatal accident data from 1983 to 1993. The results indicate that drivers with alcohol in their blood are at least seven times more likely to cause a fatal crash than sober drivers, while legally drunk drivers pose a risk 13 times greater. The study finds that the peak hours for drinking and driving are between 1:00 a.m. and 3:00 a.m., during which up to 25 percent of drivers are estimated to have been drinking. While the proportion of drinking drivers declined by approximately one-quarter over the study period, their relative crash risk remained stable. The authors also find that much of the apparent risk associated with gender or prior driving records is actually driven by differential propensities to drink and drive, though sober young drivers remain significantly more dangerous than other sober drivers. The paper calculates that the externality per mile driven by a legally drunk driver is at least 30 cents, primarily due to the roughly 3,000 innocent bystanders killed annually by drinking drivers. To internalize this externality at current arrest rates, the authors estimate a Pigouvian tax equivalent to an $8,000 fine per arrest. Furthermore, the study analyzes policy impacts, finding that higher beer taxes and stiff penalties for first-time offenders reduce the number of drinking drivers, whereas harsh penalties for repeat offenders and increased police presence do not reduce the number of drinkers but do lower the risk posed by those who do drive, likely by deterring chronic offenders or inducing more cautious behavior.
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | OpenAlex-citations | — | — | 1 | 2026-06-20 |
| archive | success | semantic_scholar | — | — | 6 | 2026-06-26 |
| extract | success | cached | — | — | 2 | 2026-06-26 |
| clean | success | clean | — | — | 1 | 2026-06-20 |
| chunk | success | chunk | — | — | 1 | 2026-06-20 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-20 |
| promote | success | — | — | — | 1 | 2026-06-20 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-26 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-20 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: crash risk outcomes