168 hours.
893 lives.
Every year.

Traffic steals 21 working days from you every year

Start with what this costs you, not the city.

Bengaluru is the second most jammed city on Earth. The average person here loses 168 hours a year just sitting in traffic . Not travelling. Sitting.

That is 21 full working days. A month of work, every year, spent staring at the back of a truck.

Everyone agrees this must be fixed. The government's answer is a road tunnel under the city, costing about ₹40,000 crore . Independent experts have questioned whether it will work at all .

So let us ask the only question that matters about any transport project.

Now turn that into a price. What does one daily rider cost you, on each option?

Every rupee you put into buses moves 39 times more people than the same rupee put into that tunnel.

And here is the same thing set against the city we actually live in. Out of Bengaluru's roughly 1 crore daily journeys, this is what each option would add:

Can you even use a road tunnel?

For most of Bengaluru, the honest answer is no.

Road tunnels do not allow two-wheelers. It is a safety and air-quality rule, applied almost everywhere in the world. Bengaluru has 87 lakh two-wheelers and only 25 lakh cars .

If you ride a two-wheeler, take an auto or catch a bus, you will pay for this tunnel every year of your life and never once travel through it.

Run the simulation yourself

Everything above is arithmetic. This is a model — and it answers the question properly: if more people took buses, how fast would Bengaluru actually move?

Traffic engineers measure road space in Passenger Car Units. A car is 1 unit, a two-wheeler 0.5, a bus 3. Divide by how many people each carries and you get the only number that matters: how much road one person uses.

Feed that into the standard speed-flow curve traffic engineers use, calibrated so that today's traffic produces Bengaluru's actual measured speed of 16.6 km/h, and you can ask what happens if more people ride.

Now untick "rebuild the bus stops" and watch the whole thing collapse.

The stop is the bottleneck, not the road

A bus is not a car. It stops every 500 metres and stands still for 20–30 seconds while people get on. Over a 10 km trip that is about 9 minutes of standing still that no amount of road-widening can remove.

Worse: a bus stop is a queue. One stopping bay can serve roughly 130 buses an hour. Today Bengaluru's busy corridors already run about 88 buses an hour past a single-bay stop — 66% full. Add 84% more buses and demand hits 147 an hour against a capacity of 92, because heavier loads also make each stop slower.

Add the buses and leave the stops alone, and the ride goes from 58 minutes to 150. The plan does not just underperform — it makes things worse.

That is not a hypothetical. It is precisely how Santiago's reform collapsed in 2007, and it is why Guangzhou built stations with several bays and a passing lane .

The surprise: you do not need everyone on a bus

Push the slider past 80% and watch the speed barely move. That is not a bug — it is how congestion works. Once the road stops being oversaturated, emptying it further does almost nothing for speed.

Going from 42% to 70% doubles the road speed. Going from 70% to 99% adds only another 6 km/h, for ₹27,000 crore more — and pushes the stops back into breakdown unless you rebuild them again.

So a 99%-bus city is the wrong target — and the model gives a second, harder reason. Drag the slider to 99% with the stops rebuilt and they break anyway: utilisation hits 100% and the ride climbs back to 135 minutes. Twin bays are not enough for that many buses; you would need four, on every stop, on every corridor.

Around 70% is where the road returns run out and where twin-bay stops still cope. Beyond that the gains keep coming in lives and clean air, but the buses start jamming each other.

I did not put that finding there. It came out of the model when I stopped pretending buses do not get in each other's way.

Why buses carry so many more

It comes down to one thing: space.

A road lane fits about 1,600 vehicles an hour, then it jams. That is physics, not Indian driving. So the only thing that matters is how many people are inside each vehicle.

Scale that to an hour and the tunnel stops looking impressive. Six lanes of tunnel move about 6,700 people an hour. One ordinary road lane, kept free for buses, moves about 6,600. Bogotá runs its bus lanes at 48,000 .

One lane of buses does the job of the whole tunnel. It costs one-hundredth as much, and the road is already there.

These prices are real, not wishful

The obvious objection is that bus plans always sound cheap on paper. So here are the actual bills other cities paid.

Four cities that already did this

None of this is theoretical. Four cities rebuilt their bus systems and published what happened. One more got it badly wrong, and that one is just as useful.

How to get more buses per rupee

Five things that decide whether ₹40,000 crore buys 10,000 buses or 6,000. None of them are technically hard. All of them are about how the contract is written.

1. Pay per kilometre, not per passenger

When an operator is paid per passenger, they race, skip quiet stops and overload the busy ones. When they are paid per kilometre driven, and the city keeps the fare money, the operator's only job is to run the scheduled service reliably. This is what fixed Seoul , and it is already how BMTC buys its electric buses.

2. Buy in bulk, and publish the price

When India's central agency aggregated bus demand across five cities into a single tender, the price it discovered was 29% cheaper than diesel buses — ₹43.49 per kilometre, published openly . Publishing the rate matters as much as winning it: once a number is public, every future tender is measured against it, and quietly paying more becomes hard.

3. Depots are the real bottleneck — solve them first

A bus with nowhere to park and charge does not run. Depot land is expensive and slow to acquire, and it is what actually delays every bus programme. Build vertically on land the city already owns rather than buying new plots, and fund depots ahead of the buses. Santiago's failure started here .

4. Run the network off the data you already collect

Every BMTC bus has GPS. Every ticket is logged. That is enough to know which stop is short of buses at 9am on a Tuesday — the question that decides whether a rush-hour bus is packed or empty. Publishing that boarding data would also let anyone check whether the schedule matches demand, instead of taking it on trust.

5. Design the stops, not just the lanes

Guangzhou's corridor moves 25,000 people an hour because its stations have several bays and a passing lane . A single-bay stop turns a fast lane into a queue. This costs almost nothing at design stage and cannot be retrofitted cheaply.

And every new road fills up again

Even if the tunnel were free, it still would not work for long.

When a road gets easier, more people drive on it, further and more often. Shops and offices open along it. People move nearby. Within a few years it is as jammed as before. This has been measured in country after country .

What ₹40,000 crore of buses actually buys

The full shopping list, at prices the Indian government has already paid.

With ₹40,000 crore you could also finish the entire stalled suburban rail — it needs ₹16,876 crore — and still have ₹23,000 crore for buses. The choice was never "tunnel or nothing".

Could we do even better?

Yes. And the biggest improvement does not cost anything extra — it saves money.

Buying buses is the obvious lever. It is not the best one. Here is what actually squeezes more out of the same money.

1. A faster bus is a free bus

A bus stuck in traffic does fewer round trips a day. Speed it up and the same bus, the same driver and the same fuel carry more people. You get extra capacity without buying anything.

This is not theory. Bengaluru measured it. The Outer Ring Road bus lane made buses 22% faster. Seoul got 26% faster across the whole city.

2. Or keep the same riders, and spend less

Turn it around. If every bus does 22% more trips, you need fewer buses to carry the same 42 lakh people.

3. So does everyone get a seat?

Mostly, yes — and this is worth being precise about, because "crowded bus" is the main reason people who can afford a car buy one.

17,000 buses carrying 90 lakh journeys works out to about 48 people boarding each bus per hour. A full-size bus has around 35 seats and room for 55. So for most of the day, on most routes, the average journey is a seated journey.

Crush loading is not an all-day problem. It happens at peak hours on maybe a fifth of routes. Fixing that is a timetable problem before it is a buying problem — and if you do need more buses on those corridors, it costs about ₹2,400 crore.

Which is, almost exactly, the ₹2,344 crore the bus lanes just saved you. You can have a seat and still come in under budget.

4. Put the buses where the people actually are

Every BMTC bus already has GPS. Every ticket is already logged. The data to answer "which stop is short of buses at 9am on a Tuesday?" already exists — it is simply not being used to set the timetable.

This is the cheapest improvement of all. It costs software and the will to change a route, not ₹40,000 crore.

When would any of this actually happen?

The most valuable thing about buses is not that they are cheap. It is that they arrive while you still care.

Bengaluru's suburban rail was approved in 2020 and promised for 2026. Today it has zero kilometres running, has lost its main contractor mid-construction, and its cost has risen from ₹15,767 crore to ₹16,876 crore before most of it broke ground .

There is a second difference, and it compounds. A tunnel is one decision made once: however many people it carries on opening day is all it will ever carry. You cannot make a finished tunnel bigger.

Buses are a decision you get to make again every year — more buses, running more often, on more lanes, reaching wherever the city has newly grown. Bengaluru is heading from 1.4 crore people to 2 crore by 2041. Only one of these two options can grow with it.

The tunnel would be jammed again by 2031, on IISc's own modelling — and we would still be repaying the loan.

1. Time

Traffic takes 168 hours a year from the average commuter here . Two things give some of that back.

If you take the bus, a bus lane is a straight speed-up. Bengaluru's own Outer Ring Road lane made buses 22% faster ; Seoul got 26% across the whole city . On a typical 10 km commute that is about 13 minutes back every day.

If you keep driving, you still gain — and this is the part most people miss. Congestion is not a straight line. Traffic engineers find that cutting peak flow by only 10 to 15% can nearly clear a jam . It is why the roads feel empty during school holidays, when barely a tenth of cars are missing.

2. Money

This one is bigger than most people expect, and it is very different depending on what you drive today.

If you ride a two-wheeler, buses will not save you money. They offer you time, safety and cleaner air instead. Anyone who tells you otherwise is selling something.

That matters, because two-wheelers are how most of Bengaluru actually travels. It is also the strongest argument for keeping fares low and using the Shakti model — the scheme that gave over 500 crore free rides and lifted women's bus travel by more than 150% .

3. Air

Bengaluru's air holds about 39 micrograms of PM2.5 per cubic metre . That squeaks under India's legal limit of 40 — and sits about eight times above what the World Health Organization considers safe.

The single biggest source is traffic. Depending on where you draw the city's boundary, vehicle exhaust and the dust thrown up by tyres account for 39% to 56% of Bengaluru's PM2.5 .

So what would 10,000 electric buses actually do to the air? Two things, and it is worth separating them.

Fewer vehicle kilometres. If total driving falls 10–15%, and transport is 39–56% of PM2.5, the city's PM2.5 drops by roughly 1.5 to 3.3 micrograms. That is a 4% to 8% improvement.

Cleaner buses. Every diesel bus replaced by an electric one removes its exhaust entirely. BMTC already runs 1,799 electric buses; this plan would take that past 10,000.

Nobody wants a car. They want a reliable way to get to work

This is the part where the research says something genuinely surprising — and it changes what we should be asking for.

Start with why people buy cars here. Study after study in developing cities finds the same thing: when public transport is unreliable, people buy vehicles as a defence, not a preference . A car is what you purchase when you cannot trust a bus to turn up.

Two things follow from that, and both matter enormously.

A car purchase is almost impossible to reverse

Once someone buys a car, they have already paid for it. The EMI runs whether they drive or not, so every trip feels nearly free even when it is not. Research also finds that car ownership carries meaning beyond travel — identity, status, a sense of having arrived . People do not sell that easily.

So every year of delay is permanent. Bengaluru adds roughly five lakh vehicles a year . Each year we wait, another five lakh households lock themselves into a vehicle they have already paid for — and become much harder to win back.

The tunnel takes five to eight years. That is 25 to 40 lakh more vehicles bought while we wait for it.

And cheaper fares are not the main lever

Here is the surprise. It is tempting to assume the answer is to make buses very cheap. The research says price helps — but service helps far more.

Transport economists measure this as elasticity: how much ridership moves when you change one thing. The published figures are consistent :

This is not an argument against cheap fares. Karnataka's own Shakti scheme proved that price matters at the extreme — making travel free lifted women's bus ridership by over 150%, and it has stayed up .

But it does tell us the order of priority. A bus that comes every five minutes and arrives on time beats a cheaper bus that comes every twenty. Which is exactly what 10,000 buses and 300 km of bus lanes buy.

Making the bus cost half a scooter

Earlier this page admitted something awkward: for a two-wheeler rider, today's ₹1,050 bus pass saves you nothing. So let us fix that, and show the whole sum.

A scooter costs about ₹7,500 to ₹10,000 a year in fuel for a 20 km daily commute . To make the bus genuinely compelling rather than merely available, it should cost half that.

The target: a bus pass at ₹400 a month — ₹4,800 a year, roughly half what a scooter burns in petrol. And a seat.

Is that affordable? Here is every rupee.

Three levers close that gap — and none of them tax bus riders

This is the part that makes the whole thing work. The money does not have to come from a new tax on ordinary people.

And the seat

A cheap bus you cannot sit on will not win over a scooter. So does the seat exist?

17,000 buses carrying 90 lakh journeys works out to about 38 boardings per bus per hour. With an average ride of 25 minutes, that means roughly 16 people on board at any moment, against about 35 seats. For most of the day, on most routes, the seat is already there.

The problem is peak hour on the busiest fifth of routes, where loads run two to three times the average. Fixing that needs about 1,870 extra buses on those corridors — roughly ₹2,400 crore.

Which is almost exactly the ₹2,344 crore that bus lanes save on the fleet. The seat pays for itself.

What else changes when a city stops idling

The speed, the lives and the air are the direct effects. These are the ones that follow — and one of them is larger than the whole cost of the plan.

The time comes back as work, or as life

At 70% bus share the model returns 84 hours a year to each commuter. Across roughly 45 lakh daily commuters that is 38 crore hours a year handed back to the city.

Value an hour at just ₹100 — well below Bengaluru's average wage — and that is ₹3,783 crore a year of time recovered. The plan's own running subsidy is ₹2,424 crore a year.

The plan pays back more than 1.5 times its annual cost in recovered time alone — before counting a single life saved or a lungful of cleaner air.

And it does not all have to become work. Ten and a half working days a year is a fortnight with your family that traffic is currently taking from you.

The air gets better than any Indian megacity has managed

PM2.5 falls from 39 to about 30 µg/m³ — a 23% cut. That takes Bengaluru comfortably under India's legal limit of 40 for the first time, with room to spare.

It is still six times the WHO guideline of 5. Buses alone cannot fix that; road dust, construction and commercial vehicles need their own answers . But a 23% cut in a decade would be among the largest air quality improvements any large Indian city has achieved.

Road injuries fall alongside deaths — from 4,052 a year to roughly 2,100. That is 1,900 people a year who do not spend a month in hospital.

And the land comes back

This is the part nobody costs, and it may be the most visible. A parked car occupies about 12 square metres of public land, free of charge, permanently.

Bengaluru's children currently play on streets shared with 1.28 crore vehicles. Fewer vehicles is not an abstraction to them. It is whether the road outside is a place you are allowed to be.

What could be done in one year

The full plan takes about four years, because buses take 12 to 24 months to build and deliver. But the biggest single win costs almost nothing and can start in weeks.

The bus lanes are the cheap part, the fast part, and the part that makes every bus you already own more useful.

Run the simulation at what that first year plausibly delivers — bus share moving from 42% to about 50% — and the model gives:

The part I had been leaving out

Every time figure on this page so far measured the time you spend inside the vehicle. A real journey is walk, then wait, then ride, then walk again — and the waiting is the part people hate most.

Transport research consistently finds that a minute spent waiting for a bus feels like two minutes riding it . Leaving the wait out does not just understate the problem; it hides the single biggest reason people buy a scooter.

So here is the honest version: a 10 km commute, measured door to door.

But here is the part that changes the politics

You do not have to switch to gain. Because the roads clear, everyone's journey gets faster on whatever they already use.

The car driver who never sets foot on a bus still gets 150 hours a year back — paid for by everyone else's bus ride. That is the argument to make to people who will never switch.

What the whole thing is worth

Three benefits, all of which can be priced. Set against the plan's ₹2,424 crore annual running cost.

And it is a jobs programme

BMTC employs roughly 32,000 people to run 7,067 buses — about 4.5 staff per bus once you count drivers on shifts, conductors, mechanics and depot staff. Running 13,000 buses needs about 58,500 people.

That is roughly 26,500 new permanent jobs — driving, conducting, maintaining. A tunnel employs people while it is being dug, and almost nobody once it opens.

What it costs to get to work, side by side

The same 20 km commute, priced every way. This is the comparison that decides whether anyone actually switches.

Today the bus costs more than a scooter, so people ride scooters. That is the entire problem, in one line.

Charging more where people can pay more

A flat fare is simple but leaves money on the table. A tiered pass raises more from those who can afford it, without pricing anyone out:

Does the premium tier pay for itself? Yes, and this is checkable. BMTC's audited cost is about ₹33 a kilometre for an ordinary bus and ₹80 for an air-conditioned Vajra . The upgrade therefore costs about ₹47 a kilometre, which works out to roughly ₹1,119 per rider per month.

Charge ₹2,000 and each premium rider contributes ₹881 a month above what they cost. At two lakh premium riders that is ₹211 crore a year flowing back to subsidise the ₹400 and ₹150 passes.

The people who would otherwise drive a car pay ₹2,000 for a better seat — and in doing so, fund the ₹400 pass for the person who would otherwise ride a scooter in the rain.

Nine facts worth sharing

Each of these is on this page with its source and its arithmetic. Take any one and post it.

Every figure links to its source on this page. If you find one that is wrong, say so — the argument is only as good as its data.

The petrol we would stop burning

Every litre not bought is money that stays in a Bengaluru household instead of leaving the city.

At 70% of trips on buses, the model removes about 50,000 lakh vehicle-kilometres a year from Bengaluru's roads. Convert that at real Indian mileage — a car does 15 km a litre, a scooter 45 — and the fuel bill that simply disappears is:

And the carbon — where I have to be careful

The petrol saved is worth 384,000 tonnes of CO₂ a year. But the new electric buses have to be charged, and India's grid still runs largely on coal.

Nobody has to take a loan

You asked the right question. Here is how the tunnel is actually being paid for.

BBMP intends to borrow ₹19,000 crore for the tunnel, with the state government standing as guarantor . It plans to borrow up to ₹39,000 crore in total for its mega projects .

The repayment comes to about ₹3,000 crore a year. The expected toll income is ₹600 crore a year.

Over thirty years, that shortfall totals roughly ₹72,000 crore — more than the tunnel costs to build.

The bus plan does not need to borrow at all

Not because buses are magic, but because of a structural difference nobody mentions: buses are divisible and a tunnel is not.

You can buy 2,000 buses this year out of the annual budget, run them, and buy 2,000 more next year. You cannot buy 40% of a tunnel and drive through it. A tunnel must be fully funded before it moves a single person — which is precisely why it requires a loan.

And once running, the plan covers its own costs:

The tunnel is a fifteen-year project in a five-year job. Buses are a five-year project that pays off in one.

How to make this happen

Not a protest. A specific, boring, winnable ask, repeated by enough people that saying no becomes the hard option.

The five asks

1. 10,000 more buses, with delivery dates in writing. Not a promise. A schedule, quarter by quarter, published where anyone can check it.
2. Build the depots first. A bus with nowhere to park does not run. Depot land is what actually delays every bus plan, and it never makes the announcement.
3. Bring back the Outer Ring Road bus lane, then add 300 km more. Under 1% of the tunnel's cost. It can start this year. Use cameras to keep cars out.
4. Compare the options in public before signing anything. Tunnel against buses against suburban rail — scored on how many people get moved, not how many vehicles.
5. Publish the tunnel's full traffic study. How many people will really use it? Who repays the loan if they do not show up? It is our money.

What one person can do in two minutes

Write to your own MLA. Not the Chief Minister — everyone writes to the Chief Minister. A letter from inside their own constituency is the one that gets counted, because it comes from someone who can vote for them.

The tool below finds your constituency, gives you your MLA's official address, and writes the letter. Change any of it you like. Then send this page to one person who will be stuck in traffic tomorrow morning.

Write the letter now →

Who actually decides

The Chief Minister and the Deputy Chief Minister, who holds the Bengaluru Development portfolio. The Greater Bengaluru Authority. B-SMILE, the company set up to build the tunnel. And the Directorate of Urban Land Transport, which is legally responsible for planning how this city moves.

None of them need convincing that traffic is bad. They need convincing that buses are the answer their voters will thank them for. That is a different argument, and it is the one worth making.

Write to your MLA in two minutes

Find your constituency, and this fills in the whole letter for you. The address below is the official one published by the Karnataka Legislative Assembly .

These are the official constituency addresses published by the Karnataka Legislative Assembly. They belong to the seat, not the person, so they keep working after an election. If a letter bounces, that is worth knowing publicly — tell us.

Check my work

An argument is only as good as its numbers. Here is where mine came from, and where I could be wrong.