I Never Thought I Would Write This…But a Private Car May be Safer than Transit for the Next Few Weeks

People are panicking about the COVID-19 outbreak.  I have read several posts on social media urging people to drive and avoid transit to reduce the spread of the disease. As of March 20, it has killed 229 people in the United States over the course of the entire outbreak so far.  In comparison, car crashes kill an average of 101 people each day.

So is the disease risk from taking transit really more dangerous than driving?  And should social distancing include practices like driving that are otherwise more dangerous than alternatives?

South Korea shows Restricting Transit is not Necessary to Contain the Virus

South Korea has probably been the country that has been most successful at containing its outbreak.  It has not restricted transit use, and their government has made assurances to the public that the transit system is safe.

Transit ridership in Seoul is down 19% as a result of the outbreak, which still corresponds to higher ridership levels than existed in the United States prior to the pandemic.

South Korea is on track for the number of infections to level off around 9000 and deaths at around 100.

Figure 1: COVID-19 in South Korea

The United States has 6.5 times the population as South Korea, so if we are as successful at containing the outbreak there would be about 650 deaths.  This would make COVID-19 much less dangerous than driving and taking transit to be the safer option. 

The United States is Probably Different

At this point it seems probable that the United States will be less successful at containing the virus than South Korea.  A major reason that South Korea has been so successful at containing their outbreak has been the most effective testing program of any country in the world.  This widespread testing has allowed the country to isolate people with the virus who have mild or no symptoms.

The United States, by contrast, has lagged behind other countries in testing, bungling its initial test implementation.

There is a serious risk that the number of infections will overwhelm the capacity of our nation’s healthcare system to be able to provide treatment.  For that reason the CDC is recommending social distancing to slow the spread of the virus. The CDC estimates that if no actions are taken to slow the spread of the virus, between 200,000 and 1.7 million Americans could be killed.  This article in Medium estimates the number of deaths without social distancing measures could be as high as 10 million. This would make the virus significantly more deadly than driving;  there were 37,133 deaths due to car crashes in 2017.

However, most viral infections occur outside of our transit system.  Does the increased infection risk of taking transit exceed the increased crash mortality risk from driving?

It turns out that the relative infection risk from transit has been little studied, and is largely uncertain. However, the risk may be lower than is commonly believed.  This study which models a hypothetical influenza epidemic in New York City estimates that 4.4% of infections would occur on the subway.  This study does not assume any reduction in ridership as a result of the outbreak. Since transit ridership is down in the United States as a result of COVID-19, this likely results in an overestimate of the transmission risk.

It is reasonable to assume that if subway ridership were restricted, at least some of these individuals would become infected in another setting. The paper did not model bus ridership transmission directly.  However, since NYC subway ridership is higher than the mode share for transit overall in the United States, it is reasonable to assume that 4.4% would be a conservatively bounding estimate for an epidemic nationally.  I would give a caveat here that the transmission of COVID-19 is less well understood than influenza, so I am not sure whether a reasonable estimate of COVID-19 subway infections would be higher or lower.

Applying the 4.4% figure to the CDC fatality estimates results in a range of 8,800 to 74,800.  If the Medium article is correct the figure would be 440,000. This compares with 37,133 deaths due to car crashes in 2017.  Since the number of choice riders who can switch from transit to riding a car is small compared with the total number of drivers, the number of additional car crash deaths from the mode switch to driving would be small compared with 37,133.  And 37,133 is the number of deaths in a year. As the Medium article describes, seven weeks is the longest that the strictest forms of social distancing will be necessary provided that our government and medical institutions use the time effectively.

Based on this, for the next few weeks those who are able should avoid transit.  Working from home, walking or biking are the safest options. If those are not available, a private vehicle is temporarily safer than transit.

Transit may still be safer than private vehicles that are ridden by a large number of people such as taxis and ride shares.

If the United States is able to bring its caseload down over the next few weeks to the point it can implement a test and isolate program similar to South Korea’s, then transit will again be significantly safer than driving.

Despite the present risks, it is important that transit service be maintained through this crisis. Transit is the only way that many people who work in essential industries are able to get to work.  This includes many staff at hospitals and medical institutions.

Can the Port Authority be Doing Anything Better?

Since the outbreak the Port Authority has improved its cleaning of vehicles.  The Port Authority is also deploying hand sanitizer at stations and taken steps to increase the physical distance between drivers and riders. These are positive steps, but more can be done.

Alon Levy reviewed international best practices, and makes the following recommendations that could be implemented by the Port Authority:

  • Provision of hand sanitizer: As noted, this is being done at stations already but it would be good for there to be a hand sanitizer dispenser on every bus.
  • Infrared sensors for temperature checks: Some transit systems screen riders at stations to prevent people with a fever from boarding.

So far there have not been any service reductions, but as ridership drops there will be a loss of farebox revenue and a temptation to reduce frequency.  If funding permits, it would be ideal to keep the buses running as this would reduce crowding and therefore the transmission risk. Houston has gone so far as to increase bus service frequency to reduce crowding and transmission risk, and bus operators ask riders to wait for the next vehicle if the bus is half full. As transit consultant Jarrett Walker explains, if frequency is cut, it is important that cuts are focused on peak frequency with off-peak frequency being maintained. 

Additionally, the relief bills being developed by congress should include funding to help transit agencies fund their improved cleaning practices and compensate them for lost fare revenue. Transit is an essential part of the transportation system in cities.

If you do Drive

If you do drive, slow down.  There is currently a lot less traffic on the road due to the social distancing measures being taken to slow the virus.  Overall this will lead to a reduction in traffic in car crash deaths and injuries. However, when there is less traffic, while the total number of crashes declines, the crashes per vehicle on the road goes up.  This is because people tend to drive faster when there is less traffic. Therefore a trip by car is more dangerous now than it was before the outbreak.

Car crash injuries put pressure on an already stressed healthcare system.  Drivers should take care to reduce their risk.

Conclusion

When I first started writing this post, I expected the conclusion to be that driving was more dangerous.  Driving is more dangerous than many things that a lot of people worry about, whether it be terrorism or mass shootings, or their children choking on something.

When I actually crunched the numbers though I realized my presupposition was wrong. That is why I take a quantitative approach, especially when evaluating risks; it informs a better choice when difficult trade-offs exist. 

Hopefully this post will be helpful for readers to stay safe during these difficult times.

Leave a Reply