Does Easier Alcohol Access Affect Road Safety? Insights from Texas Municipal Elections

Our research examines whether legalizing local alcohol sales affects traffic safety. We studied Texas, where cities can use local option elections to decide whether to permit particular types of alcohol sales. These elections are the only way a Texas jurisdiction can change its local alcohol policy: City councils and county governments cannot independently legalize or prohibit sales. Comparing cities where legalization measures narrowly passed with cities where they narrowly failed helps separate the effects of the policy from other differences between communities.

We collected records for 588 local option elections held across 466 Texas cities between 2003 and 2019. Of these elections, 527 passed and 61 failed, leading 326 cities to change their alcohol policy during the period. We combined these records with information on alcohol licenses, alcohol sales, city characteristics, and traffic crashes. We examined total crashes, crashes involving driving under the influence (DUI), and fatal crashes, with each outcome measured relative to population.

Legalization substantially increased the number of establishments selling alcohol. One year after a successful election, cities issued approximately 2.5 additional alcohol licenses per 1,000 residents compared with cities where a legalization measure failed. This increase came primarily from off-premises retailers, such as liquor stores, grocery stores, and convenience stores, where customers buy alcohol to consume elsewhere. The number of on-premises establishments, including bars and restaurants, changed relatively little.

The expansion of alcohol access also persisted over time. During the first two years following legalization, the number of licensed establishments increased by approximately 0.59 per 1,000 residents relative to a preelection average of 1.96. Nearly the entire increase came from off-premises establishments, the number of which grew by about 48 percent relative to their preelection average. The number of retail outlets remained elevated three to five years after the election, suggesting that the expansion in retail access was persistent.

Following legalization, measured alcohol sales at bars and restaurants declined, while retail store sales increased. However, these measures cover different geographic areas and cannot be compared directly. The findings are consistent with a possible shift toward off-premises purchases but do not establish that increased retail purchases replaced purchases at bars and restaurants.

Despite the substantial increase in alcohol availability, we find no meaningful evidence that legalization increased traffic crashes. Estimated effects on total crashes, DUI-related crashes, and fatal crashes were generally small and statistically indistinguishable from zero. This was true during both the first two years after an election and the three to five years that followed. We also find no meaningful evidence that legalization affected crash rates in nearby cities.

Some initial comparisons appeared to show higher total crash rates in cities where legalization measures passed. However, these cities already had higher crash rates before the elections occurred. Moreover, the apparent difference was driven by crashes not classified as alcohol-related rather than by DUI crashes. This pattern suggests that the difference reflected preexisting traffic conditions instead of an effect of expanded alcohol access.

The type of alcohol access created by legalization may help explain the results. Drinking at a bar or restaurant may require a person to drive shortly after consuming alcohol. Purchasing alcohol from a grocery, convenience, or liquor store does not create the same immediate connection between drinking and driving. Because the Texas elections mainly expanded off-premises retail access, they may have increased convenience without increasing the form of alcohol consumption most closely associated with impaired driving.

Consistent with this explanation, cities with more bars and restaurants tended to experience more crashes. One additional on-premises license per 1,000 residents was associated with approximately 1.17 additional total crashes per 1,000 residents annually, including about 0.10 additional DUI crashes. By contrast, the number of off-premises licenses had no significant relationship with any category of crashes. These comparisons are correlational and do not establish that bars and restaurants caused the additional crashes, but they reinforce the importance of distinguishing between different types of alcohol access.

Several limitations should be considered. The findings come from Texas municipalities, many of which are relatively small, and may not apply to large cities or other states. Measures of DUI crashes depend on how authorities identify and report alcohol involvement. The analysis of retail alcohol sales also relies on county-level information, making it less geographically precise than the city-level election and crash data. Additionally, the relationship between the number of bars and traffic crashes should not be interpreted as causal.

Ultimately, our findings challenge the assumption that making alcohol easier to purchase necessarily increases drunk driving. Legalization in these Texas cities expanded retail access and shifted sales away from bars and restaurants without producing measurable increases in total crashes, DUI-related crashes, or fatal crashes. The consequences of alcohol liberalization therefore appear to depend not only on whether access expands but also on where alcohol is purchased and consumed.

Note
This research brief is based on Somdeepa Das and Tong Li, “Does Easier Alcohol Access Affect Road Safety? Insights from Texas Municipal Elections,” SSRN Working Paper no. 6885959, June 2026.