The photograph is always the same: a half-finished goddess, ten arms, one eye painted, straw showing through the clay at the shoulder, a man on a ladder with a brush. It is a good photograph and it is doing the festival a disservice, because it presents the making as craft and stops there.

Kumartuli is not a craft centre that happens to produce religious objects. It is a manufacturing cluster with a four-month production season, a subcontracting system organised by process, and a product that is destroyed on delivery. Almost everything odd about the festival follows from those three facts.

Four months on, eight months off

The idol-makers of Kumartuli work to a deadline that is fixed by the lunar calendar and therefore moves: orders are placed from June and July, the frames go up in August, the clay is built up through September, the painting and dressing happen in the last fortnight, and everything must be delivered before the pujas begin in October.

That leaves roughly eight months. Some of that is spent on the other festivals — Kali Puja, Saraswati Puja, Lakshmi Puja — and some on domestic commissions, small figures for households and shops. But the shape of the year is not a smooth curve: it is a spike and a long trough, and the household's annual income is decided almost entirely inside the spike.

This is the first structural fact, and it explains the second: nobody in this trade can afford to hold inventory, because the product does not keep. A half-finished idol is not stock, it is work in progress with a hard expiry. So the cluster runs on advances, on credit from the clay and bamboo suppliers, and on a workforce that is paid in a lump at the end of the season.

The Durga Puja production chain: a four-month build feeding a one-month structure that ends in the river Jun Aug Sep Oct Nov orders frames up clay built up paint deliver workshop time about four months, income for the year pandal time built and struck bamboo frame, cloth skin, one month of life clay → straw and jute → bamboo → paint → truck → street → river Nothing is stock. The idol has an expiry date; the pandal has a lease.
Two clocks run at once: the workshop's four months, and the committee's one. The product is finished, displayed and destroyed inside a fortnight.

Why there are no factories

A cluster that makes a standardised product in high volume, once a year, for a known deadline, would normally consolidate into factories. Kumartuli did not. The reason is that the fixed assets are almost zero and the coordination is almost all social.

Consider what a workshop owns: a room, a ladder, some moulds, brushes, a yard to dry in. The bamboo and straw are bought in; the clay is dug from the river and carted in; the paint is bought; the truck is hired. Nothing is capital-intensive, so nothing pushes toward scale. What does not scale is the coordination: the work is divided by process rather than by product. One workshop builds frames. Another builds up the body. A third, often a single specialist, does the face — and the face is the part that carries the reputation, which is why the best face painters are booked a year in advance.

Subcontracting by process, in a dense quarter where everyone can see everyone else's yard, works without contracts because the information is public and the repeat game is long. That is a real industrial system; it is just not a corporate one, and it is invisible in photographs of a man painting an eye.

The customer is a committee

The buyer is not a household. In most of Kolkata it is a puja committee — a neighbourhood association that raises money from residents and local businesses, hires the idol-maker, commissions the pandal, negotiates with the police and the electricity board, and takes responsibility for a structure that will hold several thousand people a night for five days.

This changes the product. A household idol is bought for devotion. A committee idol is bought as the centrepiece of an installation that has to be photographed, competed with and reported on. That is why the pandals became design projects: bamboo scaffolding built to imitate a temple, a palace, a village, an abstract form, and dismantled after a month. The competition between committees is the engine of the whole visual culture, and it is a competition between clients, not between artists.

Tapati Guha-Thakurta's history of the contemporary pujas tracks exactly this shift: the festival's centre of gravity moving from the image to the installation, and the committee becoming a patron with a budget, a theme and a public to satisfy. Once the committee is the patron, the craft has to answer to a brief — and a brief is what turns a workshop into a supplier.

The bill that arrives afterwards

On the last day the idols go into the river. This is the part of the industry that has produced the most regulation in the last twenty years, and the reason is material.

Traditional idols were river clay and straw, painted with natural pigments, and they largely returned to the river they came from. From the 1990s, cheaper and faster materials spread: plaster of Paris for the body, because it sets quickly and can be moulded; synthetic and metallic paints for the finish, because they are brighter and last the week. Plaster does not dissolve. It settles, it forms sludge, and the paints bring heavy metals into the water.

The regulatory response has been a series of compromises: bans on plastic and on certain paints, requirements for clay bodies, designated immersion sites, and artificial tanks that are cleaned instead of the river. Each of them raises the cost of the cheapest idol, which falls hardest on the smallest workshops — the same dynamic as every environmental regulation applied to a fragmented industry. The festival is not becoming less commercial. It is becoming more expensive to do cheaply.

What the industry explains

Three things become legible once you read the pujas as a production system rather than as devotion plus decoration.

The seasonal spike explains why the craft survives at all: eight months of low income can be carried if four months are good, and a family that owns its workspace can carry them. The process-based subcontracting explains the quality and the variation: faces are the signature, so the best face painters become brands within a trade that has no brands. And the destruction of the product explains the economics: a business whose output is designed to be unusable after two weeks has no choice but to be labour-intensive, local and small.

None of that is a criticism. It is the reason the thing looks the way it does — and the reason a festival that produces a few hundred goddesses a year in a few lanes can credibly claim, as the state government did in 2019, to sit at the centre of an economy an order of magnitude larger than the idols themselves.

Sources and checks

  1. UNESCO Durga Puja in Kolkata, Representative List of the Intangible Cultural Heritage of Humanity, inscribed 2021
  2. Guha-Thakurta, Tapati In the Name of the Goddess: The Durga Pujas of Contemporary Kolkata. Primus Books, 2015
  3. Government of West Bengal Festival economy estimate submitted in support of the UNESCO nomination, 2019
  4. West Bengal Pollution Control Board Reports on idol immersion, paints and river water quality in the Hooghly
  5. Kumartuli workshop surveys Craft-census and ethnographic surveys of the idol-makers' quarter, Kolkata

Last checked 29 October 2026 Yellowsea Huang Corrections