Skip to content

Lyft Doesn’t Cause Congestion, All Vehicles Do

Restrictions targeted at the newest road users are unfair and inefficient

For years, Bob has been driving around the city to get to work, run errands, and see friends.  The traffic is sometimes bad, but given where he lives and the places he goes, it’s not practical to walk, bike, or use public transit.  Luckily for Bob, many people don’t have cars, and can’t afford taxis.  So they walk, bike, or use public transit, which keeps them (mostly) off the roads.

But then some “genius” comes along and figures out how to use smart phones to provide taxi-like ride-hailing service at lower cost and greater convenience, with hip names like Uber and Lyft. Traffic increases, and Bob’s daily travels get slower and more frustrating.  What is Bob to do? Well, for many Bobs in US cities, it appears that the answer is to hobble the technology that has allowed all those other people to get around more conveniently.UberCongestion2

After all, Bob says, he was here first, so the congestion isn’t his fault.  He just wants to make his life great again.

That’s the reasoning behind a new law in New York City — and similar moves in many cities, including San Francisco — designed to reduce congestion by limiting the number of ride-hailing vehicles. There are other cons and pros of Uber/Lyft: They spend more time looking for addresses or stopping in traffic to pick up or drop off customers. On the other hand, they spend less time looking for a parking space, stopping in traffic to try to squeeze into a parking space, or occupying a parking space.[1]

Still, the major complaint of drivers (and environmentalists) is that ride-hailing services have meant more cars on the road, and fewer people walking, biking, or using public transit. Some studies support this conclusion more than others, but it is very likely true to at least some extent.UberCongestion4

Jennifer is one of those people who doesn’t have a car, and has been walking, biking, or taking public transit.   Sometimes it’s pretty inconvenient.  Uber/Lyft has definitely improved her life.  Now, she even occasionally does things that weren’t practical before, like going to the other side of town to meet a friend for lunch. You know, the sort of things Bob has been doing for years.

So, please explain to Jennifer why she should not get to use the roads to go places with Uber/Lyft, but Bob should get to use them to drive his own car.

If you’re stuck, that’s good.  Because there really isn’t a valid explanation. Bob can twist himself into a pretzel by arguing that Uber/Lyft is making less efficient use of roads or gasoline because sometimes the driver has no passenger, or by claiming ride-hailing drivers are less skilled, but there are no credible studies showing significant differences in these dimensions.  In reality, Bob is just as much of the problem as Jennifer is.

Because the problem is not Uber/Lyft.  The problem is congestion, which is an unpriced externality created by everyone travelling by car. Bob had a good run for a while.  Only people who could afford cars and parking spaces were using the roads, so the congestion externality wasn’t too bad. Smart phones — which improved the technology for coordinating rides, and evaded some monopolistic regulation promoted by the taxi industry — have allowed a whole new population to use the roads in pretty much the same way that car owners have been doing for a century.

Just as with greenhouse gas emissions, for years we did not regulate or price congestion, because the problem just wasn’t seen as severe enough to merit government intervention.  In the last couple of decades as congestion has increased, we have taken baby steps with high-occupancy vehicle (HOV) lanes and, more recently, pricing road usage for one or two lanes on a few highways.UberCongestion1

But growing vehicle ownership, a strong economy, and now innovative ride-hailing business models have worsened congestion more quickly in many cities.  The answer isn’t to prohibit new car ownership or to throw people out of work.  And the answer isn’t to discriminate in favor of those who drive their own car over people who choose a ride-hailing service.

The answer is to deal with the problem directly.  That means figuring out how to allocate this scarce resource – space on streets and highways – fairly and efficiently among the large number of people interested in using it. Economists have long argued, and demonstrated, that pricing congestion is a really good way to help society make the best use of crowded roads.

The money collected can be used to fund the government, improve roads, or can just be distributed equitably among all people. My definition of “equitable” would be to allocate more to low-income households who would have the hardest time coming up with the road payments. But your definition may differ.UberCongestion3

If we are not willing to price road use, then we need to come up with another regulatory intervention that prevents inefficient overuse of the resource without arbitrarily favoring one model (car ownership) over another (car hire).  Unfortunately, the track record for such non-price regulation of vehicle use – such as HOV lanes and odd/even license plate driving restrictions — is not all that good.

Restricting ride-hailing isn’t the answer.  We shouldn’t make Bob’s driving life great again by undermining Jennifer’s mobility.

Severin Borenstein has no affiliation with, or financial interest in, any ride-hailing company.  In fact, he drives his own car most places he goes.  And he is still tweeting interesting energy-related articles/blogs/research most days @BorensteinS.

[1] There is also an important discussion about whether Uber/Lyft drivers are being exploited by the companies or are earning a decent income while enjoying very valuable flexibility in work hours.  Recent research by a team of highly-respected economists suggest to me the latter.

Severin Borenstein View All

Severin Borenstein is Professor of the Graduate School in the Economic Analysis and Policy Group at the Haas School of Business and Faculty Director of the Energy Institute at Haas. He received his A.B. from U.C. Berkeley and Ph.D. in Economics from M.I.T. His research focuses on the economics of renewable energy, economic policies for reducing greenhouse gases, and alternative models of retail electricity pricing. Borenstein is also a research associate of the National Bureau of Economic Research in Cambridge, MA. He served on the Board of Governors of the California Power Exchange from 1997 to 2003. During 1999-2000, he was a member of the California Attorney General's Gasoline Price Task Force. In 2012-13, he served on the Emissions Market Assessment Committee, which advised the California Air Resources Board on the operation of California’s Cap and Trade market for greenhouse gases. In 2014, he was appointed to the California Energy Commission’s Petroleum Market Advisory Committee, which he chaired from 2015 until the Committee was dissolved in 2017. From 2015-2020, he served on the Advisory Council of the Bay Area Air Quality Management District. Since 2019, he has been a member of the Governing Board of the California Independent System Operator.

16 thoughts on “Lyft Doesn’t Cause Congestion, All Vehicles Do Leave a comment

  1. Good article. Will be sharing on my blog. I know this all to well as a wrecker service owner. I live and work in Killeen TX, and this is one of the biggest problems for us. Whether we are dealing with an accident or just someone broken down on the side of the road. Thank you for the article

  2. Congestion pricing is a good idea but to say that Lyft and Uber are not the problem but congestion is is an oversimplification. Go to any intersection in downtown San Francisco and watch the traffic. So many Lyft and Uber cars! And many are without passengers. The drivers commute from outside to drive in the City because that’s where the most business is. Some regulation and limitations are necessary.

  3. The problem with these resource pricing exercises seems to be that while we may agree that pricing could/should lead to more efficient resource usage, folks immediately want to protect a large fraction of the actors from experiencing the price signal. We have seen this in GHG cap & trade, where consumers are largely insulated from seeing the costs of their higher carbon intensity decisions (and certainly don’t understand them). We also see this with these congestion pricing schemes. If roughly half of Californians are deemed low income (or disadvantaged, etc.) and you are protecting all of them from the congestion pricing mechanism, why would we expect any of those lower income residents to drive any less? You have just insulated them from the very pricing mechanism that was supposed to rationalize their usage of the resource. And higher income folks (i.e., Bob) need to get around and likely have the resources to absorb the price. So how does congestion get reduced? How does this not become just another tax on higher income Californians?

    • MSW, the solution is to separate the price signal from the income equalization mechanism. For example, SMUD provides a fixed dollar per month rebate to low income households instead of the CPUC’s method of providing 20% to 35% rate discounts. In the latter case, the CARE low income households use on average MORE power than other households. The same can be done with income-based rebates on highway tolls so the price signal isn’t diminished. The problem is the resistance that arises from much of the same constituency that advocates for relying on market-priced allocation. They refuse to directly rebate to low income households because it smells of “socialist redistribution” so instead they are willing to settle for signal suppressing price discounts for those groups. If only they would pay attention to Milton Friedman.

  4. Congestion pricing has other benefits as well. Without it, congestion is not a reliable signal of where capacity expansion is needed (Pigou-Knight-Downs and other paradoxes), but with full marginal cost pricing the traffic paradoxes disappear.

    HOTways may be moderately progressive, even before revenue distribution, to the extent that the tolls ease congestion on unpriced alternative routes with poorer drivers. The best distribution plan may be a simple income-tax credit, the transparency of which would enhance political feasibility (a major constraint).

    • Jim – I don’t understand either of your points – but to focus on the second. I think that toll lanes (I assume that’s what you mean by Hotway) might ease congestion ONLY when the lanes are added to the existing freeway. In the case of the lanes on I-580, as I recall, one lane was added in each direction, which in short order became a toll-lane, but then a second lane was converted from ‘free’ to toll. Depending upon where one is on I-580, that’s either a 1 out of 6 reduction or 1 out of 7 in terms of ‘free lanes’. Hard to see how that relieves much traffic congestion — at rush hours the ‘free’ lanes are barely moving.

      • On Jim’s second point, instead of introducing pricing complexity through a inverse means test, he is suggesting a straight forward income tax credit (it also could be means tested). Since most households use the freeways, even as transit riders, it will effectively reach the same targeted population. If we’re concerned about rebating to those using a local road network, we can screen the tax credit eligibility by zip code.

        • Richard,
          Thanks for these insights. Let me try again to make three points:
          1. By rendering MB=MC, setting tolls equal to marginal social cost facilitates supply-side traffic management as well as solving demand-side issues. This is often explained via the traffic paradoxes, e.g., but the point is more general. (One can make the same point for power networks, although electricity is more well-behaved than drivers and more reliably modeled.)
          2. The congestion-pricing literature typically assumes that you can price all roads. So the discussion of HOTways needs a separate literature to deal with only some roads being priced. If you’re adding a HOTway (e.g. as proposed in Honolulu as an alternative to the current rail debacle), then users get consumer surplus and non-users may get less congestion than in the counterfactual. As Rich points out, changing a freeway to a tollway is a different matter.
          3. By distributing toll revenues in a predetermined (so voters don’t think the politicians will waste them) and such that the vast majority of users have transparently higher net benefits, road-pricing is more likely to be politically feasible.

  5. Isn’t this – on its face – another example of a “classical” limits to growth problem – combined with our usual inability to engage is regional or metropolitan-area planning? I’m not sure what is meant by “inefficient overuse” as opposed to overuse that is efficient? Its still overuse – in this case traffic congestion, not to mention the attendant increase in air pollution, poor land use decision making (parking vs open space?), etc. Cities, left to their own devices, have shown themselves to be inherently unable to think beyond their city boundaries – my favorite example in the SF Bay Area is Dublin, which appears to have annexed almost all of what was once open land north of I-580 from I-680 to almost Livermore (I don’t live in either city, so I’m a little unclear on those details). In the past 10 or 15 years, that whole region has become covered with housing – requiring new freeway interchanges, shopping centers, etc. – all of which means greater use of automobiles. One consequence is the not unexpected outcome of more traffic on the interstates – even though I-580 was widened a few years ago to accommodate addition lanes. Now, the two “fast” lanes east of I-680 have congestion pricing – which means that the folks that can least afford if (and have had to move further away from higher and higher rental prices) pay an even greater penalty of increased traffic congestion and longer commute times being stuck in the slower (but “free”) lanes.

    I dunno – maybe there is some way to rebate congestion pricing fees via an inverse means test – but given the perversity of our current politics, I’m not holding my breath. In the meantime, traffic congestion is but one measure of ‘growth’ in a resource-constrained environment (or what should be constrained). In the case of Dublin, where are they getting the water to support additional population (speaking of constrained resources)?

  6. This is an interesting problem in property rights. Let’s make a further assumption that Bob is a resident of the city where he’s experiencing more congestion. As a renter/homeowner, he has paid taxes to build the roads in the current configuration that assumed individual vehicle ownership. Let’s then assume Jennifer has just moved to that city. Does Bob have a property right in the existing road system that he has paid for up to this point and was designed with his driving mode in mind? And does that mean that Jennifer has to “buy” her way into the local road network? When we can clearly define property rights, how to price this access would also be clear. But this network externality, as with other public good externalties, create complications without clear solutions.

    This reminds me of a midterm question in introductory microeconomics at Berkeley taught by Steve Goldman. The question was to design a pricing system for an academic computer network. It turned out that there was no clean answer, and the question was intended to see students’ reasoning process. We face the same dilemma when pricing road networks. Who are the “beneficiaries” who should be paying? Are there vested property rights? How should network expansion costs be allocated?

%d bloggers like this: