The ferry sets the price of grain

The first boat leaves at five, in the dark, and it is already over its limit. A bicycle is lashed to the rail, two sacks of rice are wedged against the engine housing, and fifteen people are sitting on a bench built for eight. Nobody has a ticket. The man collecting fares works his way along the gunwale with a cloth bag, charging by category rather than by distance: a person, a person with a bicycle, a person with a motorcycle, a sack, a goat.

This is the ordinary morning of a river crossing in the Gangetic plain, and it is the least photographed piece of infrastructure in the country. It has no building, no schedule board and usually no name. Yet it sets the price of grain, the length of a school day and whether a sick person can come home the same evening.

Two prices for one sack

Take a sack of paddy grown on the far bank. Its farm-gate price is set by the best buyer who can physically reach the field, and that buyer is rarely the one offering the highest price — it is the one willing to cross. A trader on the near bank can pay less, because the farmer's alternative is to carry the sack down to the ghat himself, pay per sack, wait for a crossing, and then arrive at a market where he is a stranger with one sack and no place to store it overnight.

Across the river, the same sack sells for more. The gap is not freight. Freight on a boat is cheap per kilo. The gap is the cost of the three things a boat cannot guarantee: a departure, a space, and a return. A truck can be hired at any hour and will wait while you negotiate. A boat leaves when it is full, carries what fits, and stops at dusk. Everything that cannot be priced exactly — the queue, the wait, the risk of being stranded — gets priced into the grain instead, as a discount on one side and a premium on the other.

This is why the price gap between two banks can be larger than the price gap between two districts a hundred kilometres apart on the same side. A district road, however bad, is continuous. It does not stop at sunset. You can push a cart along it, or walk it, or pay someone to. The river is not continuous, and the crossing is not a road with water on it — it is a service, owned by someone, with hours and a capacity, and that service is the bottleneck through which every economic decision on the far bank has to pass.

What a boat can do that a bridge cannot

There is one advantage the boat keeps, and it is not a sentimental one. A river in this plain changes course. The Kosi has moved tens of kilometres east over the past two centuries, and in August 2008 it shifted again after a breach upstream in Nepal, putting three districts under water and displacing around three million people. A bridge built to span the 2008 channel may be spanning a sandbank by 2050. A boat does not have this problem. It moves to wherever the water is, which is why the ferry survives not as a relic but as the only crossing that can be re-routed without capital.

That same flexibility explains a detail that visitors find strange: the ferry often runs on a different alignment in February than in August. In the dry months the channel narrows and splits around sandbars, and the boat works a shorter crossing between two temporary ghats. In the flood months it runs a longer, slower line over ground that is normally farmland. The fare changes with the crossing. Nothing is announced; everyone knows.

The bridge arrives, and rewrites the map

When a bridge does come, the effect is not gradual. Prices on the two banks converge within a couple of seasons, because the sack can now cross at night, in a truck, in any quantity. Land values on the far bank rise — abruptly, and mostly for owners who were not the ones carrying sacks. The boatmen lose a trade that their families had held for generations and which had no formal title, so there is nothing to compensate. And a market town that existed only because it sat at the landing stage begins to empty, while a new one forms at the bridgehead, at a different point on the map, with different owners.

This is why the location of a bridge is a more consequential decision than its design. A bridge placed at the district headquarters serves the administration. A bridge placed where the boats already run serves the economy that the boats had already built. The two are rarely the same point, because the boats went where the current allowed and the headquarters went where the colonial revenue map said.

Why the metric misses it

Road density is the standard measure of rural connectivity, and it is a poor one in country like this. A district can report excellent road density and still have a third of its villages on the wrong side of a river with no crossing after four in the afternoon. The number that predicts how a household fares is not kilometres of road per square kilometre. It is the distance to the nearest crossing that works in the monsoon, and the time of the last return.

Which is why the most useful question you can ask about a village here is not how far it is from a town. It is: what time does the last boat come back. Everything else — whether the children go to secondary school, whether a woman in labour reaches a hospital, whether a farmer sells at the farm gate or at the mandi — follows from a departure time chalked on a board that nobody outside the village will ever read.