Did Uber and Lyft Reduce Drunk Driving? Research and NC Data

DWI Data & National Comparisons

Did Uber and Lyft reduce drunk driving? The honest answer is “sometimes, somewhat, and not everywhere.” Peer-reviewed studies are split. Some find drops in alcohol-related crashes or DUI arrests after ridesharing arrived, others find no change, and one large study finds overall traffic deaths went up. North Carolina’s own numbers show no clear rideshare effect. Alcohol-impaired deaths rose through the 2010s, spiked during the pandemic, and only recently started to fall.

The short answer: did Uber and Lyft reduce drunk driving?

The research on whether Uber and Lyft reduced drunk driving is genuinely mixed. Several studies find modest reductions in alcohol-related crashes, fatalities or DUI arrests, while other well-designed studies find no average effect. The best reading is that ridesharing can help in some places and for some people, but it has not solved impaired driving anywhere.

The idea is intuitive. If a cheap, quick ride home is a few taps away, fewer people should drive after drinking. Rideshare companies made exactly that argument as they expanded across the country after 2012. But a good argument is not the same as evidence. Researchers have spent a decade testing it with crash databases, arrest records and the staggered dates when Uber and Lyft entered different cities. The results depend heavily on the place studied, the years covered, and whether the researchers measured fatal crashes, all crashes, arrests or total traffic deaths.

This article summarizes what the major studies actually found, with links so you can read them yourself. It then sets those findings beside North Carolina’s federal crash data and the state’s own crash counts for Mecklenburg, Iredell, Cabarrus and Stanly counties. It closes with the practical part: how to get home safely around Charlotte, Lake Norman and Stanly County, and why the cost of a ride is trivial next to the true cost of a DWI in North Carolina.

What the peer-reviewed research found

The strongest positive finding comes from California, where UberX’s arrival was linked to a 3.6% to 5.6% drop in alcohol-related motor vehicle deaths. National studies of the largest U.S. metro areas found no average effect on alcohol-involved fatalities. Here are the main studies, in the order they were published.

Studies that found reductions

Greenwood and Wattal (2017), California. Writing in MIS Quarterly, Brad Greenwood and Sunil Wattal compared California markets before and after UberX arrived, using data from 2009 to 2014. They reported a 3.6% to 5.6% decrease in the rate of alcohol-related motor vehicle homicides per quarter after UberX entered a market. The higher-priced UberBlack service showed only intermittent effects, which suggests price matters. People substitute a ride for driving when the ride is cheap. (MIS Quarterly 41(1):163-187)

Dills and Mulholland (2018), U.S. counties. Economists Angela Dills and Sean Mulholland combined federal fatal-crash data with FBI arrest data. In their basic model, each additional quarter that Uber was available was associated with a 1.6% decline in the fatal crash rate, and each additional month with a 0.8% decline in DUI arrest rates. An alternate model put the fatal-crash effect at 0.7% per quarter. Most specifications showed declines in DUI arrests. (Southern Economic Journal 84(4):965-991)

Morrison and colleagues (2018), four cities. Researchers from the University of Pennsylvania used a natural experiment: cities where Uber started, stopped and restarted service. In Portland, Oregon, Uber’s return was associated with a 61.8% relative drop in the alcohol-involved crash rate, about 3.1 fewer alcohol-involved crashes per week, with no change in all injury crashes. In Reno there was no significant change. San Antonio’s results were inconsistent and based on limited data. The authors stressed that effects vary by city and time. (American Journal of Epidemiology 187(2):224-232)

Studies that found no effect, or harm

Brazil and Kirk (2016), 100 largest metro areas. Noli Brazil and David Kirk tracked when Uber arrived in metropolitan counties from 2005 to 2014. They found that Uber’s deployment was not associated with later traffic deaths, including drunk-driving deaths and deaths on weekends and holidays. (American Journal of Epidemiology 184(3):192-198)

Brazil and Kirk (2020), follow-up through 2017. With more years of data (2009 to 2017), the same authors again found that Uber was not associated with changes in total, alcohol-involved, or weekend and holiday traffic deaths in aggregate. The effect varied by place, and Uber was linked to more traffic deaths in dense urban counties. (PLoS ONE 15(9):e0238744)

Barrios, Hochberg and Yi (2023), U.S. cities. This study looked at all traffic deaths, not just alcohol-related ones. The arrival of ridehailing was associated with roughly a 3% increase in traffic fatalities and fatal accidents, for both vehicle occupants and pedestrians. The authors attribute it to more cars on the road: more vehicle miles, more congestion and more new car registrations. The effect was larger in big cities and in places that used public transit and carpools more before ridehailing arrived. (Journal of Operations Management 69(5):823-855; NBER Working Paper 26783)

What the reviews say

Two efforts have tried to pull the literature together. A 2022 review presented by Columbia epidemiologist Christopher Morrison at the Research Society on Alcoholism meeting looked at 20 studies (14 peer-reviewed, 6 unpublished). It reported that 17 of them linked ridesharing to lower alcohol-impaired driving or alcohol-related crashes. That was a conference abstract, not a full peer-reviewed paper. A 2025 systematic review and practice guideline from the Eastern Association for the Surgery of Trauma, published in the Journal of Trauma and Acute Care Surgery, found 15 qualifying studies. Eleven showed benefit for specific outcomes and one showed harm (more total crashes). The authors rated the evidence low quality because every study was observational, but concluded that ridesharing “may have a valuable role to play” in reducing crash harm.

Why the studies disagree

The studies disagree because they measure different outcomes, in different places, over different years. Ridesharing can reduce one kind of harm (drunk drivers on the road) while adding another (more total miles driven). Both effects can be real at once.

Three issues explain most of the conflict. First, outcome choice. Fatal crashes are rare, so city-level fatality counts bounce around and small effects are hard to detect. Studies of all alcohol-involved crashes or DUI arrests have more data to work with, but arrests also reflect police staffing and priorities. Second, place. A rideshare network with many drivers and short wait times near a dense bar district is a real alternative to driving. In a rural county, a 30-minute wait at 1:30 a.m. may not be. Brazil and Kirk’s 2020 finding that effects differ by county is a warning against assuming one national answer applies to your town.

Third, everything else changing at the same time. The 2010s brought smartphones, distracted driving, bigger vehicles, new enforcement campaigns, and, after 2020, a pandemic that emptied roads and then filled them with faster, riskier driving. Isolating the effect of one app against all of that is hard. The Morrison team’s approach of watching what happens when service stops and restarts in the same city is clever precisely because it holds many of those other factors steady. Even then, the results varied by city.

A ride app only prevents a DWI if someone decides to use it before the keys come out. The research tells us the option helps. It cannot make the decision for you.Adkins & Purkey

Rideshare in Charlotte and North Carolina: a timeline

Ridesharing reached Charlotte between 2012 and 2013, spread to the Lake Norman towns by 2015, and has been regulated statewide since North Carolina’s transportation network company law took effect on October 1, 2015.

1. Sept. 2012Uber’s Charlotte blog says the service first appeared in Charlotte during the Democratic National Convention, which the city hosted that September.
2. Summer 2013Lyft expands to Charlotte, according to WFAE.
3. Sept. 2013WFAE reports that UberX and Lyft have newly “landed in Charlotte,” catching city officials by surprise.
4. Feb. 13, 2014Lyft expands to most of the metro area and becomes the first major rideshare service to pick up passengers at Charlotte Douglas International Airport.
5. June 5, 2015Uber expands to Concord, Cornelius, Davidson and Huntersville. Lyft already served Huntersville and Concord.
6. Oct. 1, 2015Session Law 2015-237 (Senate Bill 541) takes effect, creating a statewide permit system for transportation network companies.
7. 2020-2021The pandemic. North Carolina’s alcohol-impaired driving deaths jump from 386 (2019) to 454 (2020) and 466 (2021).
8. 2023-2025Alcohol-related deaths begin to fall. NCDOT counts 349 alcohol-related deaths in 2025, the lowest in its 2020-2025 series.

That timeline matters for reading the data. Rideshare was widely available in Charlotte by 2014 and in the north Mecklenburg towns by mid-2015. If ridesharing had a large, immediate effect on drunk driving in North Carolina, you might expect alcohol-impaired deaths to fall after 2014 or 2015. They did not.

North Carolina alcohol-impaired crash and fatality trends

North Carolina’s alcohol-impaired driving deaths, as counted by NHTSA’s Fatality Analysis Reporting System, rose from 368 in 2013 to a peak of 466 in 2021, then eased to 415 in 2023 before rising again to 449 in 2024. Alcohol-impaired deaths have stayed between roughly 26% and 30% of all traffic deaths in the state throughout.

NHTSA defines an alcohol-impaired-driving fatality as a death in a crash where at least one driver or motorcycle operator had a blood alcohol concentration of 0.08 or more. The count includes everyone killed in those crashes: the impaired driver, passengers, other motorists and pedestrians. The chart below uses NHTSA FARS figures as reported in North Carolina’s FY2023 Highway Safety Plan for 2013 through 2019, and NHTSA’s annual Alcohol-Impaired Driving fact sheets for 2020 through 2024.

Read the chart in three phases. From 2013 to 2019, the years when rideshare spread across Charlotte and then the suburbs, alcohol-impaired deaths in North Carolina did not fall. They rose from 368 to a 2016 high of 428, then hovered around 400. Total traffic deaths rose faster, so the alcohol-impaired share dipped slightly, from about 28.5% in 2013 to about 26.5% in 2019. Then came the pandemic. Alcohol-impaired deaths jumped 18% in a single year, from 386 in 2019 to 454 in 2020, even though people were driving less. They peaked at 466 in 2021. Since then the numbers have come down, though 2024 (449) was still well above the pre-pandemic level.

For the United States as a whole, NHTSA counted 11,904 alcohol-impaired-driving deaths in 2024, 30% of 39,254 traffic deaths. North Carolina’s 28% share sits slightly below that national figure. Our article on North Carolina alcohol-impaired driving statistics breaks down the 2022 numbers.

NCDOT’s own count, and the local picture

NCDOT publishes a separate series based on crash reports (Form DMV-349) where officers coded alcohol involvement. Because the definition differs from FARS, the numbers do not match exactly, but the trend is similar. Statewide, NCDOT counted 440 alcohol-related deaths in 2020, 442 in 2021, 479 in 2022, 395 in 2023, 391 in 2024 and 349 in 2025. Alcohol-related crashes of every severity stayed between about 11,200 and 12,400 a year over the same period. NCDOT notes that 2020 reflects pandemic-era traffic and may not be typical.

Mecklenburg County, with the region’s densest rideshare market, still averaged nearly 1,000 alcohol-related crashes and about 49 alcohol-related deaths a year from 2020 to 2024. Its alcohol-related deaths fell from 58 in 2022 to 40 in 2024 and 43 in 2025. Stanly County’s numbers are small enough that one crash changes the picture, so year-to-year swings there mean little.

The legal trap: waiting for your ride in the driver’s seat

In North Carolina, sitting behind the wheel with the engine running can be “driving” for DWI purposes, even if the car never moves. People who do the right thing by calling a ride can still be charged if they wait for it in the driver’s seat with the motor on.

G.S. 20-138.1 makes it a crime to drive while impaired on a street, highway or public vehicular area, and parking lots usually count. North Carolina courts read “drive” broadly. A person in actual physical control of a vehicle that is moving or has its engine running is driving. It does not matter that you only started the car to run the heater or air conditioning while you waited. Our parked-car case study and our article on State v. Overocker show how these cases arise.

Safe-ride options around Charlotte, Lake Norman and Stanly County

Around Charlotte and the north Mecklenburg towns, rideshare, taxis and CATS transit are the main options. In Stanly County and other rural areas, plan the ride before the first drink, because late-night service is thinner.

• Rideshare apps. Uber and Lyft have served Charlotte since 2013 and Huntersville, Cornelius, Davidson and Concord since at least 2015. Check the app before you go out, and look at wait times at the hour you expect to leave, not when you arrive.
• Taxis and car services. Licensed taxis and sedan services operate in Charlotte and around the airport. Save a number in your phone in case an app is surging or has no drivers nearby.
• CATS buses and the LYNX Blue Line. Charlotte Area Transit System serves Charlotte and parts of the region. Hours vary by route and day, and Blue Line work around South End this fall includes scheduled single-tracking. Check CATS schedules or call 704-336-7433 before relying on transit for a late trip.
• Designated driver. Still the oldest and cheapest solution. Choose the driver before anyone starts drinking, and make sure that person really does not drink.
• Stay over. A hotel near the restaurant, a friend’s couch, or the guest room at the lake house costs less than a tow, and it removes the decision entirely.
• Stanly County. Stanly County’s public transportation service, SCUSA (704-986-3790), runs subscription and demand-response routes. Call ahead to ask about hours and booking. Rideshare and taxi coverage in Albemarle, Locust, Oakboro and Norwood can be limited late at night, so arrange a ride or a sober driver in advance.

The cost of a ride vs. the cost of a DWI

Even an expensive ride home costs a tiny fraction of a DWI. In our hypothetical first-offense example, court, DMV and program costs alone exceed $1,300, and with a five-year insurance surcharge the total can pass $20,000.

Rideshare prices change by the minute with distance, time and demand, so check the app for your own trip. Suppose, hypothetically, that a late-night ride from uptown Charlotte to Huntersville costs $45 and the ride back for your car the next morning costs another $45. That $90 compares with a list of DWI costs that starts before any conviction: a tow and storage bill, a $100 fee to end the 30-day civil revocation, and days without a license.

After a conviction, the official costs pile up: $293 in court costs, a fine set by your sentencing level, a $100 assessment and $160 school or treatment, about $167 in DMV restoration fees, and often an ignition interlock at $50 to $92 a month. The largest cost is insurance. A DWI is 12 points under the Safe Driver Incentive Plan, a 340% surcharge. For convictions on or after July 1, 2025, that surcharge can last five policy years (see insurance points after a DWI). Then add lost work, a lawyer, and the risk of hurting someone. The full itemized breakdown, with a cost estimator, is in our guide to the cost of a DWI in NC.

What this means if you were charged anyway

If you were charged with DWI, the availability of a ride home is not a legal defense, but the facts of how and where you were found matter a great deal. Whether the engine was running, where the keys were, whether you were in a public vehicular area, and how long you had been there can all bear on whether the State can prove you were driving.

Write down the timeline while you remember it: when you stopped drinking, when you called or tried to call a ride, what the app showed, where you were sitting, and where the keys were. Save screenshots of ride requests and messages. Those details can corroborate your account. Then deal with the license side quickly. The 30-day civil revocation starts right away, and a limited privilege requires a substance abuse assessment. Drivers under 21 face separate rules; see under-21 alcohol and driving. Our list of first steps after a Lake Norman DWI walks through the first week.

Frequently asked questions

General North Carolina legal information as of November 2026, not legal advice for any specific case. Reading this article does not create an attorney-client relationship. Attorney advertising. Ride and cost comparisons are hypothetical illustrations.