Odd-Even Traffic Rules: A Simple Idea With a Complicated Reality
When a city struggles with traffic congestion and severe air pollution, policymakers often look for interventions that can be implemented quickly. One of the most visible approaches is the odd-even traffic rule, where vehicles are permitted to operate on particular days depending on the last digit of their registration number.
The concept appears simple: cars ending in an odd number operate on odd-numbered dates, while cars ending in an even number operate on even-numbered dates. On paper, this can immediately reduce the number of eligible private cars on the road.
But does reducing the number of cars automatically reduce congestion and pollution?
The experience of Delhi, along with examples from Beijing, Mexico City, Bogotá and other cities, suggests that the answer is more complicated. Odd-even restrictions can influence traffic volumes and travel behaviour, but their environmental impact depends heavily on public transportation, exemptions, vehicle substitution, meteorological conditions and compliance.
The issue also demonstrates an important lesson for analytics: a policy that looks fair mathematically may not necessarily produce a fair or effective real-world outcome.
Where Did the Odd-Even Concept Come From?
License-plate-based driving restrictions are not new. Cities have used variations of vehicle rationing for decades as a way of controlling congestion, pollution and demand for limited road space.
Historical examples include Caracas, which introduced a restriction in 1979, Santiago in 1986 and Mexico City in 1989. Similar approaches later appeared in São Paulo and several Colombian cities, including Bogotá and Medellín. China subsequently became one of the most prominent users of license-plate restrictions, particularly in Beijing.
The underlying principle is known as road-space rationing or license-plate rationing.
Instead of charging drivers to use congested roads, the government restricts when certain vehicles can use them.
This creates a straightforward policy mechanism:
Fewer eligible vehicles → potentially lower traffic volume → potentially lower congestion and emissions.
However, the chain is not guaranteed. Drivers may respond by buying a second vehicle, using another mode of transport, travelling at a different time or shifting their trip to another day.
That behavioural response is where the analytics becomes particularly important.
Delhi's Odd-Even Experiment
Delhi brought the concept into the national spotlight in India when the government introduced its first odd-even experiment from January 1 to January 15, 2016. A second phase followed from April 15 to April 30, 2016. Under the scheme, private cars with odd-numbered plates were permitted on odd dates and even-numbered cars on even dates, subject to exemptions.
The policy was introduced against the backdrop of extremely poor winter air quality in Delhi.
The January 2016 experiment was not simply a traffic-management exercise. It became a large-scale real-world experiment involving millions of vehicles, commuters, roads and public-transport users.
Researchers subsequently studied:
Traffic volumes
Vehicle occupancy
Travel speeds
PM2.5 and PM10 concentrations
Other pollutants
Public-transport usage
Carpooling behaviour
Meteorological conditions
Driver responses
This makes Delhi particularly interesting from an analytics perspective because the policy generated an enormous natural dataset.
An Important Calendar Insight
The original analysis behind this topic also illustrates how seemingly small mathematical details can influence policy analysis.
A common assumption is that odd and even dates occur equally often. They do not necessarily do so in a calendar year.
For example, in a normal 365-day calendar year, the distribution of odd and even calendar dates is not equal because months have different lengths. In a leap year, the difference changes again.
However, this should not be interpreted as meaning that odd-numbered vehicles automatically receive a meaningful real-world advantage. Actual access depends on the precise policy rules, including Sundays, exemptions and the period during which the scheme operates.
This is an important distinction between calendar mathematics and policy outcomes.
A mathematically interesting imbalance may exist, but it does not necessarily translate into a comparable advantage for drivers.
What Did the Delhi Experiment Actually Show?
This is where the story becomes much more interesting.
Different studies reached somewhat different conclusions because they examined different datasets, locations, time periods and methodologies.
A study published in Transportation Research Record examined the January and April 2016 experiments and found that car flow rates declined by less than 20%, while traffic from two-wheelers, buses and autorickshaws increased. The researchers also found little evidence that car owners substantially adopted car sharing. They concluded that the scheme did not produce a measurable reduction in PM2.5.
Another study examining Delhi's air pollution found that the policy's impact was difficult to separate from weather conditions. During winter, low wind speeds and poor atmospheric dispersion can cause pollutants to remain trapped over the city.
At the same time, research focusing on traffic corridors found evidence of reductions in particulate matter during the first phase. One study reported an average reduction of approximately 5.73% in PM2.5 and 4.70% in PM1.0 across three monitored corridors.
These apparently conflicting findings demonstrate why policy evaluation cannot rely on a single metric.
A reduction in traffic at a particular road does not necessarily mean a proportional reduction in city-wide pollution.
Case Study 1: Beijing
Beijing provides one of the most frequently cited examples of license-plate restrictions.
The city introduced strict odd-even restrictions around the 2008 Olympic Games, when authorities needed to manage traffic and improve environmental conditions. Research found evidence of reductions in congestion and mobile-source pollution during the restriction period. Beijing subsequently continued with less restrictive forms of license-plate-based driving restrictions.
But Beijing also demonstrates an important limitation.
If restrictions remain in place for a long period, drivers can adapt.
For example, households may purchase additional vehicles with different registration numbers. A policy that initially reduces vehicle use can therefore become less effective as people change their behaviour.
This is known as behavioural adaptation.
Case Study 2: Mexico City and Latin America
Mexico City introduced a license-plate restriction in 1989. Similar policies subsequently appeared across Latin America, including São Paulo and Bogotá.
These examples demonstrate that license-plate restrictions are particularly attractive to governments because they are relatively easy to communicate and enforce.
However, they also reveal a recurring challenge:
Restricting vehicle ownership is not the same as reducing vehicle travel.
If a household owns another vehicle, switches to another mode or changes the timing of its trips, the expected environmental benefits can decline.
Therefore, the success of an odd-even system depends on the entire transportation ecosystem rather than the rule itself.
Case Study 3: Delhi as a Data Analytics Experiment
Delhi's experience is arguably more valuable as a case study in policy analytics than as proof that odd-even rules are either successful or unsuccessful.
Researchers have examined traffic data alongside pollution measurements, meteorological conditions and travel behaviour.
One study using statistical techniques investigated the effects of the odd-even policy while controlling for factors such as weather, fuel prices, agricultural burning and other potential influences.
This is critical because pollution is a multi-variable problem.
Consider the simplified relationship:
Air Pollution = Vehicle Emissions + Industrial Emissions + Dust + Biomass Burning + Weather + Regional Pollution + Other Sources
Reducing one component does not guarantee a large reduction in the final outcome.
If wind speeds are low and pollution from neighbouring regions is transported into the city, removing some private cars may have only a modest effect on overall air quality.
Real-Life Applications Beyond Pollution
Odd-even restrictions can also be used for purposes beyond air pollution.
1. Managing Traffic Congestion
During major events, governments can temporarily limit vehicle access to reduce congestion around stadiums, exhibition centres or city centres.
2. Emergency Traffic Management
During severe pollution episodes, natural disasters or infrastructure failures, authorities can use vehicle restrictions to reduce pressure on critical roads.
3. Managing Road Capacity
When a major highway or bridge is undergoing construction, license-plate restrictions can temporarily reduce demand.
4. Event-Based Mobility
Large international events can generate extraordinary transportation demand. Temporary vehicle rationing can help cities manage this peak demand.
5. Data-Driven Urban Planning
Perhaps the most valuable application is not the restriction itself but the data it generates.
Cities can compare:
Traffic before the restriction
Traffic during the restriction
Traffic after the restriction
Average vehicle occupancy
Public-transport demand
Pollution levels
Travel times
Road speeds
Geographic variation
This transforms a policy into a measurable experiment.
What Would a Better Analytics Model Look Like?
A modern evaluation should go beyond simply counting odd and even vehicles.
A comprehensive model could combine:
Vehicle Data + Traffic Data + Pollution Data + Weather Data + Public Transport Data + Geographic Data + Behavioural Data
For example, an analytics dashboard could monitor:
MetricBefore PolicyDuring PolicyAfter Policy
MetricBefore PolicyDuring PolicyAfter Policy
Private vehicle volume
Baseline
↓
Baseline
Average speed
Baseline
↑/↓
Baseline
Travel time
Baseline
↓/↑
Baseline
PM2.5
Baseline
↓/↑
Baseline
Public transport usage
Baseline
↑
Baseline
Car occupancy
Baseline
↑/↓
Baseline
The objective should not be to prove that the policy works.
The objective should be to determine under what conditions it works, for whom, where and by how much.
The Bigger Lesson for Businesses and Governments
The odd-even rule offers a powerful lesson in analytical thinking.
A policy can appear logical at first glance but behave differently when implemented in the real world.
The original question was essentially:
“If we reduce the number of cars, will we reduce pollution?”
A better analytical question is:
“What happens to total transportation demand when certain vehicles are restricted?”
That second question accounts for substitution, behavioural change, public transportation, weather and other sources of emissions.
This is the difference between descriptive analytics and decision analytics.
Descriptive analytics tells us what happened.
Predictive analytics helps estimate what may happen next.
Prescriptive analytics asks:
“What should policymakers do differently?”
Conclusion: From Odd and Even Numbers to Smarter Cities
The odd-even traffic rule is much more than a simple calendar-based driving restriction. It is an example of how governments can use a relatively simple rule to influence complex urban behaviour.
Delhi's experience shows that the effectiveness of such a policy cannot be judged simply by counting vehicles removed from the road. Different studies have reported different effects on traffic and particulate pollution, partly because air quality depends on many factors beyond private-car traffic.
The strongest lesson is therefore not that odd-even rules are either “good” or “bad.”
It is that transportation policies need continuous measurement, experimentation and evidence-based adjustment.
For modern cities, the future is likely to involve a combination of congestion pricing, better public transportation, cleaner vehicles, intelligent traffic management, emissions controls and real-time data analytics rather than dependence on a single restriction.
The odd-even rule may have started with a simple idea—odd cars on odd days and even cars on even days—but its real value lies in what it teaches us about data, human behaviour and the complexity of managing a modern city.
The ultimate objective should not be to create an odd-even city. It should be to create a smarter, cleaner and more efficient one.
This article was originally published on Perceptive Analytics. At Perceptive Analytics our mission is "to enable businesses to unlock value in data." For over 20 years, we've partnered with more than 100 clients — from Fortune 500 companies to mid-sized firms — to solve complex data analytics challenges. Our services include AI Consulting Services and Power BI Consulting Services in Phoenix, turning data into strategic insight. We would love to talk to you. Do reach out to us.
Top comments (0)