The tea economy, in figures

TeaHQ holds 1,737 measured figures about tea production, trade, prices and certification. None of them was estimated, interpolated or converted between currencies here; each one names the publisher it came from, the year it refers to, and the definition that publisher uses.

If a number is not on this site, the most likely reason is that we could not retrieve it from the body that publishes it. There is no field in our data model for a figure nobody published, and no page here will ever show one.

Almost every wrong tea statistic in circulation is this one mistake. FAO’s crop database reports the primary product— green leaf, as it comes off the bush. Every national tea board, and FAO’s own Intergovernmental Group on Tea, reports made tea — the finished dried product that is actually sold. Roughly four and a half kilograms of leaf make one kilogram of tea.

Every table below is on one basis and says which. This is the only place on the page where both appear together, because this is the only place where the comparison is the point.

Cells are marked with the publisher’s own confidence flag where it is not a plain official return: e estimated by the publishing agency, i imputed by a receiving agency, x carried in from an external organisation, p provisional, m the output of a balance-sheet model. An unmarked cell is an official figure reported by the country. A series that changes flag part-way through changes method part-way through, and the step in the line is the method, not the harvest.

Three things are visible here that no single-year figure shows. The first is how differently the same six decades treated different countries: on FAO’s numbers Kenya’s crop is more than two hundred times its 1961 level and China’s more than a hundred and sixty times, while Sri Lanka’s has multiplied by roughly seven. Sri Lanka did not shrink. It simply did not do what the others did.

The second is the shape of China’s recent growth. Its crop roughly doubles between 2010 and 2024 — and its recorded area under tea more than doubles over the same span. That is expansion, not yield.

The third is a warning rather than a finding. Kenya’s series changes flag in 2021, from estimated by FAO to official, reported by Kenya, and the level shifts at the same moment. That step is a change of method. Reading it as a harvest event is exactly the mistake the flag column is there to stop.

Cells are marked with the publisher’s own confidence flag where it is not a plain official return: e estimated by the publishing agency, i imputed by a receiving agency, x carried in from an external organisation, p provisional, m the output of a balance-sheet model. An unmarked cell is an official figure reported by the country. A series that changes flag part-way through changes method part-way through, and the step in the line is the method, not the harvest.

In 1961 Sri Lanka shipped sixteen times as much tea as Kenya. It shipped more through the 1970s and 1980s and most of the 1990s. Kenya passed it for the first time in 1993, fell back, and has been ahead without interruption since 2005 — by 2023 it exported more than twice as much, on a Sri Lankan volume barely changed from 1961. The 2024 Kenyan cell is imputed: FAO fills it at 466,670 tonnes, while the Tea Board of Kenya’s own report for the same year says 594,500. Both numbers are on this site. Neither is marked correct.

The World Bank publishes its tea auction quotations twice in one workbook: once in the dollars of the day, and once in constant 2010 dollars, deflated by its Manufactures Unit Value index. Read the first column and tea has got dearer. Read the second and it has not.

FAO’s Intergovernmental Group on Tea states the same thing in words: “international prices of tea in real terms, that is adjusted for inflation, have been declining for the past four decades, reflecting a tendency for technical improvements to increase yields and production faster than demand.” A series that started in 1990 could not be used to check a claim about four decades, which is why this one starts where the World Bank’s record does. In 1960 the three-auction average was USD 1.03 a kilogram in the money of the day, and USD 5.38 in 2010 dollars. In 2025 it was USD 2.91 and USD 2.51. Nominally it nearly tripled. In real terms it more than halved. Both sentences are true, and only one of them is about whether growing tea pays.

The picture most readers carry is of an estate: a plantation, a manager, a workforce. That has not been the shape of the industry for a long time. The Tea Board of India splits every growing region into Big Growers — estates with their own factories — and Small Growers, who sell green leaf to a bought-leaf factory. Both figures below are published tonnages; the share is arithmetic on them, shown so you can check it.

The spread is the interesting part. In the Terai, four fifths of the tea comes from small growers; in Darjeeling, in the hills immediately above it, almost none does. Kenya’s own returns tell a similar story from a different direction: of the tea its factories made in 2024, the estates account for under a quarter, with the rest coming through smallholder co-operative factories, independent bought-leaf factories and the state Nyayo zones.

You will often read that smallholders grow 60 percent of the world’s tea. FAO states it, its Intergovernmental Group on Tea repeats it, and the working-group paper that repeats it footnotes a document from 2014. None of them attaches a reference year or a method, so TeaHQ does not record it as a statistic. The table above is what a publisher actually counted, in a year it named.

The most common question about the tea economy is what share of the price on a supermarket shelf ends up with the person who picked the leaf. It has a clean-sounding answer in a lot of places. It does not have one here, and the reason is worth more than a percentage would be.

The most careful public synthesis we found reports, within a single page: that wholesalers and retailers take about 45 percent of the final retail price; that estates can obtain margins of about 15 percent of the total retail price; that in China fresh tea retains about 20 percent of its total value added, with blending and refining taking 15 percent; and that in Viet Nam farmers capture about 42 percent of net value added. Those are figures on at least three different denominators — retail price, total value added, net value added — from four separate studies, in three countries, spread across nine years. They do not average into anything.

What settles it is that the body with the mandate does not have the number either. In January 2024 the FAO Intergovernmental Group on Tea’s Working Group on Smallholders put a proposal to the Group’s twenty-fifth session: to build an open-access platform, with participation “from all producing and consuming countries as well as trade associations and research institutes”, that would calculate what a sustainable consumer price should be. The intergovernmental body responsible for the commodity is proposing to construct the figure, because it does not exist.

We will publish a grower’s share of a retail price when somebody publishes one we can retrieve, for a stated year, on a stated denominator. Until then the honest output is this paragraph, and a plausible percentage here would be the most quotable and least true sentence on the site.

How TeaHQ handles sources