Dhaka’s traffic congestion is hardly an undiscovered problem. Successive governments have built flyovers, overbridges and underpasses, restricted daytime truck movement, proposed moving inter-district bus terminals out of congested areas, expanded mass transit, cleared roads and footpaths, experimented with signals, and periodically intensified enforcement. Yet gridlock remains an everyday feature of life in the capital. That raises a more interesting question than what causes the jam: why has congestion remained so persistent despite so many attempts to reduce it?
A 2017 World Bank assessment reported that average driving speed in Dhaka had fallen from 21 kilometres per hour a decade earlier to less than seven, while congestion consumed about 3.2 million working hours every day. The conventional explanation is too many people, too many vehicles and too few roads. All three matter, but they are not a sufficient diagnosis. Congestion is the system-level outcome of road scarcity, mixed vehicles, intersections, buses, pedestrians, parking, enforcement, synchronised schedules, institutional fragmentation and, above all, the incentives that govern how people use scarce road space.
Two lenses, one framework
An economist sees scarcity, incentives and externalities. A physicist sees density, velocity, bottlenecks and flow instability. Through both lenses, Dhaka’s traffic jam is a complex-system problem, neither purely engineering nor purely economic. This is where an orthogonal perspective and matrix paradigm can help. The orthogonal perspective first decomposes congestion into distinct analytical axes so that no single explanation dominates the diagnosis. The Matrix then reassembles them and examines how they interact. Understanding each axis separately is necessary. Solving them separately is not sufficient.
Road space: scarcity or misallocation?
Dhaka certainly has limited road space, but the relevant quantity is not kilometres of asphalt. It is usable capacity at the place and time it is needed. A four-lane road does not deliver four lanes when one is occupied by parked vehicles, vendors, waiting buses or informal passenger pickup, and clearing it once does not help if encroachment returns the moment enforcement leaves. Babubazar Bridge illustrates the point. Recent reporting documented stalls, unauthorised stands, passenger pickup and parked buses on this important route. The bridge exists. Part of its effective capacity does not.
So, the policy question is not only how much new road Dhaka should build but also how much capacity it can recover from roads already built. Dhaka could designate zero-obstruction mobility corridors at peak hours, prohibiting parking, prolonged stopping and commercial encroachment, and concentrate scarce enforcement where small gains produce large network benefits. Road-clearing drives are events, while congestion is continuous; predictable enforcement changes incentives.
Intersections: where traffic multiplies itself
Here the first interaction appears. Recovered road space is worth little if it feeds into a blocked junction, and junctions are where congestion becomes nonlinear. A vehicle that enters a crowded intersection without room to exit stays put when the signal changes, blocks the perpendicular flow, and may back traffic into the next junction. A local mistake propagates through the network. This is why the small patches of asphalt where major roads cross deserve disproportionate attention. A strict “do not enter unless you can exit” rule at critical junctions could yield benefits far greater than its simplicity suggests, provided buses do not stop near those junctions and parking, pickup and roadside commerce are kept at a distance.
Technology can help, but Dhaka’s history warns against technological solutionism. Artificial intelligence may optimise a signal; it cannot clear an intersection physically blocked by vehicles. Recent AI-supported enforcement at selected intersections has reportedly improved compliance with lights and zebra crossings, suggesting that technology works when it changes the expected consequences of behaviour.
The bus paradox
The clearest case of incentives overriding design is the bus. Buses ought to be among Dhaka’s most efficient users of road space, since they carry many passengers in little room. Yet poorly aligned incentives can turn them into congestion generators.
A 2021 field-level study found that commuters lost roughly one working hour a day to congestion, and respondents named passenger pickup and drop-off at arbitrary locations as the most important perceived cause.
Why do buses behave this way? Under daily-contract or target-revenue arrangements, drivers are rewarded for capturing passengers. Racing rivals and stopping outside designated places may be bad for the system but individually rational for someone whose income depends on reaching the passenger first.
The disorder is partly the predictable outcome of the compensation system, not merely poor driver discipline. Bus reform therefore means more than painting stops on the roadside. Route rationalisation and company-based operations should be paired with fixed salaries or compensation tied to reliability, safety and schedule performance.
A bus should earn money by moving people efficiently, not by winning a roadside contest. Designated stops must mean designated stops; otherwise, every metre of road becomes a bus station, and one person’s convenience becomes a delay for thousands. The troubled BRT experience shows that good transportation technology inside bad incentives can still produce bad outcomes.
Rickshaws and the physics of different speeds
Rickshaws raise a different issue. They cannot be wished away. They provide livelihoods, cheap mobility and essential first- and last-mile service. The difficulty is physical, not moral: vehicles of sharply different speeds force repeated overtaking and merging, whose disturbances intensify as density nears capacity. Banning them from one corridor can simply push traffic onto surrounding roads. The answer is functional separation where feasible: rickshaws on neighbourhood roads and as feeders to transit, with access to selected arterials managed at peak hours. The aim is compatibility of flows, not preference for one class of traveller over another.
Parking and pedestrians
Some of the scarcest public land in the metropolis serves as free or underpriced storage for private vehicles. A parked car creates a delay far beyond its own footprint. Where unavoidable, parking should be designated, priced and barred at peak hours on critical corridors. Pedestrian infrastructure teaches a similar lesson. If climbing an overbridge costs far more effort than crossing at street level, and the expected penalty is negligible, many people will cross below. Crossings must be convenient, accessible and backed by enforcement. Sidewalks must also work: when vendors and parked vehicles push pedestrians into the roadway, a pedestrian problem becomes a vehicle-capacity problem.
Enforcement: why rules on paper do not move traffic
Running through all of these is enforcement. Dhaka probably does not lack traffic rules. It lacks certainty that they will be enforced at a particular time and place. If an illegal manoeuvre saves ten minutes and the chance of penalty is negligible, violating the rule looks rational, and multiplied across hundreds of thousands of road users, individually rational decisions add up to collectively irrational congestion.
Enforcement must also be symmetrical. Non-essential trips and private or family use of publicly funded vehicles at peak hours impose road-space costs with no public benefit. Vehicle pools, trip logs and limits on non-essential peak movement would let government demonstrate the discipline it expects of citizens. Scarce road space does not become less scarce because the vehicle occupying it carries an official number plate.
Time: why everybody leaves home at once
So far the problem has been spatial. It is also temporal. Schools, government offices, banks and businesses run on overlapping schedules, while ministries, hospitals and major institutions cluster in a limited part of the city. Centralisation creates tidal commuting flows; synchronised schedules compress them into narrow windows. A network that handles a given number of trips spread over several hours may become unstable when too many arrive within one. Economists call this peak-load demand.
Staggering the opening and closing times of schools, government offices and selected private institutions could redistribute traffic across time without eliminating a necessary journey or building a meter of road. It should be targeted, introduced experimentally and judged by traffic data. Over the longer term, directing new offices and educational facilities toward satellite centres such as Purbachal and Savar could weaken the pull of the old core. Staggering addresses when people travel; decentralisation addresses why so many must travel to the same places.
Governance: too many decision makers
Perhaps the hardest dimension is institutional, because it determines whether any of the above gets done. Congestion ignores bureaucratic boundaries. One agency administers the road, another enforces traffic law, another regulates buses, municipal decisions shape sidewalks, and land use is governed through yet another process. Each can perform its narrow task while the system fails; the interactions among institutions may be what is failing.
Dhaka therefore needs an operational traffic-management architecture that integrates real-time information, coordinates interventions and assigns accountability for system performance. Police should be judged partly by network outcomes such as intersection clearance and queue length, not only by violations counted. A city has one traffic system. Its management should behave as though it does.
Why sensible policies keep disappointing
Much of Dhaka’s apparently irrational traffic behaviour becomes understandable from the road user’s seat. The bus driver stops suddenly because another passenger means more revenue. The motorist parks illegally because legal parking is distant. The pedestrian crosses below the overbridge because climbing costs effort. Each responds to incentives.
This explains why individually sensible policies disappoint. Clear illegal parking and effective capacity rises, but if buses occupy the recovered space while waiting for passengers, much of the gain vanishes. Regulate buses but leave the next junction blocked, and the queue returns. Clear the junction while occupied; sidewalks push pedestrians into the road, and another disturbance appears. A policy can succeed on its own axis and still disappoint the system.
There is also a bias worth confronting. Large projects are visible: flyovers and rail lines can be photographed and inaugurated. Operational discipline is not glamorous, yet in a congested system, mundane management may yield returns comparable to far costlier additions of capacity. Dhaka’s problem has not been excessive infrastructure investment. It has been capital investment insufficiently complemented by operational discipline.
The new master plan: necessary but not sufficient
The newly unveiled Strategic Transport Plan 2026–2045, with its eight metro lines, two BRT lines, three ring roads, six motorways and 21 multimodal hubs, shows the scale of official ambition. These investments may well be necessary, and they reinforce the argument here. A ring road’s exits can become bottlenecks, and a metro station without feeder transport and clear pedestrian access underperforms. Infrastructure creates capacity. The Matrix determines how much of it becomes usable mobility.
Experiment, measure, then expand
Seen separately, Dhaka appears to have a road problem, a bus problem, a parking problem, an enforcement problem and a governance problem. Seen from a bird’s-eye perspective, they are interacting components of one system, so the right philosophy is not to search for the single solution but for the combination of interventions whose interactions reinforce rather than neutralise one another.
Dhaka need not wait for another generation of megaprojects nor turn the whole city into a laboratory overnight. Select three to five chronically congested corridors as integrated demonstration zones. Recover capacity through zero-obstruction enforcement at peak hours. Protect junctions with clearance rules. Require designated bus stops and weaken roadside passenger competition through better compensation. Manage mixed vehicle flows selectively, enforce peak-hour parking limits, keep sidewalks usable, test staggered schedules and hold government vehicles to the same discipline as private ones. And place each corridor under one clearly identifiable authority.
Above all, measure everything before and after: average travel speed, intersection throughput, queue length, bus travel time, pedestrian crossing time and the variability of journey times. Successes can be expanded; failures modified or abandoned. The approach is falsifiable, so a critic need not accept the theory. Let the traffic data decide. It is less glamorous than inaugurating a flyover and may be considerably cheaper.
The bird’s-eye view
No serious analysis should promise to eliminate congestion from one of the world’s densest metropolitan areas. The realistic goal is more modest and more valuable: turn chronic gridlock into manageable congestion, raise effective traffic velocity and make journey times predictable. An econo-physicist looking at Dhaka from above sees not badly behaved vehicles but millions of individually motivated decisions interacting inside a network of finite capacity. Physics tells us what happens when density overwhelms flow, economics why people keep adding to that density, and institutional analysis why well-meant interventions fail to reinforce one another.
Dhaka’s traffic jam is a Matrix problem. It requires a Matrix solution. And perhaps the fastest road Dhaka can build is the one it has already built but has not yet learned to use efficiently.
_________________________________________________
The writer is a former physicist and nuclear engineer at the BAEC and Professor Emeritus of Economics at Eastern Michigan University, USA. He can be reached [email protected]
For all latest news, follow the Daily Sun's Google News channel.
