Coffee & Economics
I enjoy trying new coffee and cafés, and looking for a good one has become part of how I travel. That habit turned into Know your Grounds, a map of specialty cafés around the world that I built and keep adding to.
Coffee also runs through the parts of economics I work on. It is grown in a small number of tropical countries and drunk mostly in rich ones, so every cup depends on a long international supply chain, and climate change puts that chain under strain at the farm, at the port and on the shipping routes. A frost in Brazil or a drought at the Panama Canal feeds through to the world price, which makes coffee a clean example of a supply shock moving along a chain of international trade. And the price of the finished cup differs across countries in ways that line up with standard facts in international macroeconomics about prices, wages and exchange rates. The page takes these in turn: where the money in a cup goes and where climate hits the chain, and then a small index of how many minutes of work it takes to buy the cappuccino you just ordered — remember this number the next time you tip your local barista.
Know your Grounds
A global map of specialty cafés you can trust
A café earns its place by naming the roaster it buys from, by brewing pour-over and syphon by hand alongside espresso, and by being independent, since chains are excluded outright. Cafés the coffee press has written up score higher, while crowd star-ratings count for nothing, and the method behind the score is published in full so anyone can see why a café sits where it does. I also add the ones I have particularly enjoyed on my own travels.
Why climate matters for the coffee supply chain
Coffee reaches the consumer through a long supply chain, and most of what a cup costs is added a long way from the farm. Taxes on coffee differ so much between countries that they swamp everything else, so the figure below strips them out. What is left divides roughly into thirds: a third to the farmer, half to roasting and retail together, and an eighth to moving the coffee. Four-fifths of the farmer’s third goes on growing the crop, which leaves an income of about a twelfth of the price.
Notes and sources
Euros per kilogram of German national-brand ground coffee, 2021, shown as shares of the price before tax. Taxes are taken out because a flat German excise of €2.19 a kilogram is a fact about Germany rather than about coffee: the same levy is 31% of a private-label ground pack and 6% of a capsule. One origin-to-Germany chain for one product, not a world average.
Source: BASIC et al. (2024), The Grounds for Sharing, commissioned by the Global Coffee Platform, IDH and Solidaridad, as reproduced in Coffee Barometer 2026, Figure 7, p.59. Division of the pre-tax price is the author’s calculation from their published euro figures.
Climate presses on the early stages hardest, and its effect thins as it travels. A world price that roughly doubled between early 2024 and February 2025 left the price of a cup of coffee in a café close to where it started. The green coffee in a $5 drink was worth about 17 cents at the top of that run, against nearer 7 cents at more ordinary prices.
Where climate and disruption hit the chain
The farmer
Frost and drought in Brazil have preceded every large spike in the world price since the 1970s. Robusta, the hardier half of the crop, loses about 14% of its yield for every degree above a mean of 20.5°C, an optimum several degrees below the range the industry had assumed (Kath and others, Global Change Biology, 2020). Coffee leaf rust compounds this, since its worst epidemics have come in low-price years when growers cut back on management (Avelino and others, Food Security, 2015). Rain then damages coffee that has already been grown and picked, at the point where it is drying. Above 750mm in the harvest window, with mean minimum temperatures over 22°C, the chance of above-average defects passes three in four (Kath and others, Climate Risk Management, 2021). In November 2025 more than 1,100mm fell on Vietnam’s Central Highlands in five days, against a November norm of 104mm, while part of the harvest lay on the drying patios (International Coffee Organization, Coffee Market Report, November 2025).
Export
In September 2024, 84% of container ships at Santos were delayed and 2.2 million bags of coffee sat backlogged. Cecafé attributes the congestion to container infrastructure that has not kept pace with the volumes moving through it (Cecafé, October 2024).
Trade and shipping
After the driest October since 1950, 41% below normal, the Panama Canal cut daily booking slots from about 36 into the low twenties through 2023 and 2024, and one transit slot went at auction for $4 million (Panama Canal Authority, Advisory A-48-2023). Red Sea diversions then pushed four to five million bags into coffee not yet imported at destination, tightening European supply while every harvest stayed where it was (International Coffee Organization, Coffee Market Report, December 2024).
Roasting
Green coffee is about 68% of what one of the world’s largest roasters spends on product, and energy is too small a line to report separately (JDE Peet’s, full-year results 2024). The exposure that matters at this stage is carbon pricing, with the European Union’s second emissions trading system now due in 2028.
Retail and cafés
At Starbucks, running the stores costs about half of revenue while every product sold costs under a third (Form 10-K, financial year 2025). Wages, rent and milk carry that half, and a frost in Minas Gerais leaves them where they were.
The cappuccino index
The cappuccino index was created by James Hoffmann, a former World Barista Champion, author and authority on all things coffee. He built it from a crowdsourced survey in which baristas report their hourly wage and the price of a small cappuccino where they work. The index is the local price divided by the local wage of the person who made the drink. Since both sides are in the same currency, the exchange rate cancels and none of the comparison depends on how a currency is valued. What is left is a price measured in working time, which is a real wage upside down — the lower the number, the more coffee an hour behind the machine buys. A cappuccino costs a barista in Australia about ten minutes of their own working time, and a barista in India a little under three hours.
Because the survey records prices and wages in local currency, the same cups can be set against two standard textbook facts in international macroeconomics. The first is that the prices of goods and services that do not trade across borders tend to rise with a country’s income, which is the Balassa–Samuelson effect. The second is that purchasing power parity holds poorly, so that the same good can cost very different amounts once prices are converted at market exchange rates, which the Economist’s Big Mac index has shown for decades. Hoffmann explains how he built the index, and what went wrong while he was building it, at rather more length than this page does.
Hoffmann on the index → The full data →
Barista time and income
Notes and sources
Minutes of barista work needed to buy one small cappuccino, against real GDP per capita, 2024. Each point is a country. The dashed line is an ordinary least squares fit through all 36 points in logs, with a slope of −1.02 and an R² of 0.69.
Sources: index from James Hoffmann’s Cappuccino Index 2026, a crowdsourced survey of 2,594 reported pairs of barista hourly wage and cappuccino price, restricted here to the 36 countries answered at least ten times. GDP per capita from World Bank World Development Indicators, series NY.GDP.PCAP.PP.KD, PPP in constant 2021 international dollars (CC BY 4.0). The fitted line is the author’s calculation.
The fit is surprisingly strong, especially for an index built on cappuccinos! The slope in logs is about minus one. That is, doubling income per head roughly halves the barista time a cappuccino costs. Countries below the line get their coffee cheaply for the income they have, and the three furthest below are Australia, Italy and New Zealand, which will surprise nobody who has ordered a coffee in any of them. Singapore, Turkey and Malaysia sit furthest above.
The downward slope might look like it contradicts Balassa–Samuelson, but it is consistent with it once the index is split into its two halves. A cappuccino is close to a pure nontradable, since even at the top of the 2024–25 run the green coffee in a $5 drink was worth under 4 per cent of its price, and what the customer pays for is mostly local labour, rent and milk. As the textbook predicts, its price rises with income: across the same 36 countries, a 10 per cent higher income per head goes with a cappuccino about 3.5 per cent dearer. Barista wages, however, rise much faster, by about 13.6 per cent, and because the index is price divided by wage, it falls. Richer countries pay more for the cup but pay the person making it more again, which is the productivity story behind Balassa–Samuelson seen from the other side of the counter.
A new coffee-based measure of exchange-rate valuation
The second textbook fact concerns exchange rates. If purchasing power parity held, a cappuccino converted into dollars at the market exchange rate would cost the same everywhere. Where it does not, the gap between a country’s dollar price and the American price measures how far its currency is from parity. The Economist has done this with the Big Mac since 1986. Dividing a country’s Big Mac price in local currency by the American price in dollars gives an implied exchange rate, and comparing that with the market rate says whether the currency is overvalued or undervalued against the dollar. I ran the same computation using the cappuccino prices in the survey, which gives a new measure I call the Cappuccino Exchange Rate Index. It broadly agrees with the Big Mac index: across the 34 countries in both, the correlation is 0.69, and in seven countries out of ten the two agree on whether the currency is overvalued or undervalued.
Notes and sources
Over- and undervaluation against the US dollar, in per cent, implied by the price of a cappuccino (the Cappuccino XR index) and by the price of a Big Mac (the Big Mac index). Each point is one of the 34 countries in both indices; a country on the dashed line is valued the same way by the two. The United States is omitted, since it is zero on both by construction.
Sources: cappuccino prices from James Hoffmann’s Cappuccino Index 2026, the same survey as above, in local currency and restricted to the 36 countries answered at least ten times; Big Mac prices and market exchange rates from The Economist’s Big Mac index, July 2026, github.com/TheEconomist/big-mac-data. Implied rates, and Big Mac valuations for individual euro-area countries, are the author’s calculation.
Arguably, the disagreements between the two indices are driven by the goods underlying them. The nominal exchange rate cannot explain these disagreements, since the euro trades at one rate in Rome and in Berlin, and yet the Cappuccino Exchange Rate Index has it 57 per cent undervalued against the dollar in Italy and 9 per cent undervalued in Germany. Some of the gap between the indices is likely measurement, because the cappuccino price is an average over a crowdsourced survey of cafés while the Big Mac price is a single item on a standard menu. Local preferences over the goods themselves are probably part of it too. In the United States a cappuccino is closer to a luxury good, whereas in Italy it is part of daily life for far more people, drunk standing at an espresso bar, in a smaller cup and at a price that custom keeps low (€1.85 in the survey). Because every implied exchange rate is computed relative to the American price, the first of these helps explain why cappuccino prices imply that most currencies are more undervalued against the dollar than Big Mac prices do, with an average undervaluation of 16 per cent against 1 per cent. The second helps explain Italy, where the euro is 57 per cent undervalued on cappuccino prices and 21 per cent overvalued on Big Mac prices.