Breaking the gridlock: La Paz (Bolivia) edition
Policy contagion and the likely effects of a traffic-reduction policy
Introduction
In March 2023, the La Paz Council extended the area of license plate-based driving restrictions. This policy restricts cars with license plates ending in a certain digit from entering the city centre on specific weekdays. For example, cars with a last digit of 1 or 2 can enter the city centre on all weekdays but Monday.
The stated goals of this policy are to reduce congestion and pollution. These problems are hardly unique to La Paz. So, why implement this policy instead of taking an alternative course of action like charging vehicles for access, improving public transport, doing nothing, or encouraging people to work from home?
Policy contagion
The role of mimicry
One explanation for choosing this policy is that similar cities already have it. Copying similar cities’ policies is an adaptive strategy for two reasons.
The first reason is information gain. The fact that a policy was implemented in other places signals that the policy is feasible, both politically and logistically. Ideally, it also signals that the policy works (i.e., achieves its stated objectives).
The second reason is political. It’s easier to justify a policy that exists in similar cities than one that doesn’t exist at all or exists only in dissimilar cities. In case the policy is typical and doesn’t work (or even backfires), the politician can always find safety in numbers, arguing that he acted like everybody else around him. Conversely, if his policy is atypical and fails, he would be denounced as reckless and irresponsible. The message to politicians: it’s fine to fail as long as the cause of your failure is a widely adopted dysfunctional policy.
Reference groups and how policies spread
People can only copy what they are aware of, which may or may not include members of the relevant reference group. The most natural reference group for La Paz is other Latin American cities. The shared language, geographical proximity, and similar economic situation make it more likely that these cities and La Paz will be similar on the policy-relevant dimensions. Importantly, La Paz residents probably also accept these cities as the reference group.
If people copy (within some reasonable bounds) whatever others are doing, a policy can be widely adopted in a whole region even when there is little or no evidence that it works. To see why, imagine that the current policy is a laissez-faire approach to congestion. A city in the reference group (e.g., Bogotá) decides to implement license plate-based driving restrictions. The next city in line, when faced with public pressure to do something about congestion, finds it easier to justify adopting the same policy than adopting an alternative policy. As more and more cities adopt the same policy, its gravitational pull becomes stronger, and any subsequent city will find it increasingly hard to choose an alternative policy. The result of this dynamic is that license plate-based restrictions become very common throughout Latin America, while, for instance, congestion charging is virtually non-existent.
Uniform policies are dangerous because they disincentivize experimentation and lead to complacency. After all, if this policy is adopted in dozens of cities, people assume it must work. But does it?
Policy effects
Effects of license plate-based restrictions
As far as I can tell, there is little research on the effects of this type of policy, especially if we consider how widely it is used. A recent review study found mixed evidence. The variability in the quality, design, and methodology of the included studies makes any conclusions highly questionable. To give you a flavour of just one source of variation, the included studies look at measures as diverse as increased television viewership, hospital admissions, pollution, car trips, travel times, traffic speeds, number of cars without catalytic converters, and gasoline consumption.
Effects in La Paz
I couldn’t find any studies that analysed the policy’s effects in La Paz. For lack of a better alternative, I will evaluate the policy based on its intended effects.
The mayor announced that 27 000 fewer cars would enter the city centre on weekdays. I must admit I was positively surprised to see a number. This is superior to the usual strategy of setting a very broad, pleasant-sounding yet impossible-to-evaluate objective like making the city better for its residents.
Before we delve into the details, a quick note on the measure (i.e., cars entering the city centre). I’m not sure the number of cars that enter the city centre is a good proxy for what we care about, which is the amount of traffic in the city centre. For example, a taxi could count as one entry, yet circulate in the city centre all day long, while another person may enter the city centre twice but spend only 2 minutes there. A more fundamental concern, perhaps, is whether the 27 000 reduction refers to the number of cars entering the city centre or some other measure that looks similar on the surface but is fundamentally quite different. Unfortunately, the municipal press releases mentioning this number don’t use terms very consistently. So, going forward, I will be using the term traffic reduction, keeping in mind that the mayor may be referring to a different measure.
Hidden in another press release I found a reference to a presentation by the mayor, in which he had said that 134 000 cars enter the city centre every weekday. The intended traffic reduction of 27 000 corresponds to a 20% reduction (after some rounding). I suspect the mayor arrived at this number by assuming two things. First, that everyone was using his or her car for 5 weekdays. Second, that the person will only use it for 4 weekdays once the policy takes effect. This would indeed lead to a 20% traffic reduction. Unfortunately for policymakers, people have the annoying tendency to change their behaviour when their incentives change.
Responding to incentives
Before I offer some thoughts on how people might change their behaviour, the first thing to note is that many of the current drivers won’t be affected by this policy. Imagine you were driving to the city centre three times a week before the policy was implemented. After the change, you may act as before if your restricted day happens to be one of the two days you weren’t going to drive to the city centre anyway. So, even if people don’t change their behaviour, the intended 20% traffic reduction is almost certainly an overestimate. But people do change their behaviour. Let’s see some examples.
Driving the same amount, but on different days or at different times than before
If people are not allowed to drive to the city centre on Wednesday between 7 am and 8 pm, they may go on Tuesday at any time, or on Wednesday outside the restricted hours.
Driving the same amount, but with a different vehicle
People can use a family member’s or a friend’s car on a restricted day. In extreme cases, they can purchase a second car.
Ignoring the policy or stealing number plates
Enforcing the policy might be difficult. If the probability of being caught is low, some people may be willing to risk a fine. Alternatively, they may think of the fine as the price to pay to enter the city centre on their restricted day. Although a remote possibility, some people might resort to stealing license plates that end in a different digit than theirs.
The temporary reduction in traffic encourages people to switch from public transport to driving
If there is a temporary reduction in traffic, driving becomes more attractive to public transport users. Some of them might switch to driving.
Conclusion
Some of these possibilities are very real, while others are more remote. But they all shrink the effect of the policy towards zero. In theory, the effect of the policy could be outside the 0 to 20% range. For instance, the effect could be larger than 20% if the policy has no unintended consequences, and on top of that, may expose all-time car users to public transport, who realise that taking the bus is not that bad after all and decide to use it more frequently in the future.
Realistically, the likely reduction in traffic is in the 0 to 20% range. If I had to make a prediction, I would expect an effect somewhere between 0 and 5%, assuming that all else remains the same. Having said that, we should not forget that we considered only the benefit side of the equation, and the policy’s costs can easily outweigh its expected benefits.
Unfortunately, we are unlikely to learn if my prediction will prove correct. This would require deciding in advance how success and failure will be measured, tracking progress (or lack thereof), and being transparent about the findings. This is non-glamorous, hard work. The easier, well-trodden path is to call the traffic situation a crisis and promise iron-fisted policies that will solve the crisis once and for all. Let’s hope eventually more and more politicians will be incentivised to choose the harder path.