The CRS Robusta Report

The last market close included in this issue. The issue date is when it was published; the close date is what the numbers are actually made of. They are usually a day apart.

The trading days behind every figure on this page: dates on which both exchanges published certified stock, given as a count and a date range. A date on which only one exchange reported is left out, because a comparison needs both sides.

The archive is young. Where a window is as long as the archive itself, the page says so rather than implying history it does not have.

The International Coffee Organization's published gap between the New York arabica futures price and the London robusta futures price for the month, in US cents per pound. It is the ICO's figure, cited as theirs; this report does not calculate a futures arbitrage of its own.

A wide gap means arabica is expensive relative to robusta on the futures markets. It is a monthly figure, so it will not move with the daily stock numbers beside it.

The quantity of arabica coffee sitting in ICE-approved warehouses that has been graded and passed, and is therefore deliverable against a Coffee C futures contract. Reported here in thousands of 60-kg bags. It is not all the arabica in store — only the portion the exchange has certified as deliverable.

A falling number means certified arabica is leaving the exchange system, either drawn into consumption or decertified. It is a measure of exchange-deliverable supply, not of world supply, so it can fall sharply in a year with a large crop.

The same thing for robusta: coffee in ICE-approved warehouses, graded and deliverable against the Robusta contract. ICE reports robusta in lots of 10 tonnes; this row converts to the same unit as the arabica row above — thousands of 60-kg bags, at 166.67 bags to the lot — so the two can be compared directly.

Compare it with the arabica row only in the converted unit shown. A lot count and a bag count are not comparable numbers, and differencing them raw produces a plausible-looking figure that means nothing.

Certified arabica divided by certified robusta, both in bags-equivalent. One number for how the exchange-deliverable pool is split between the two coffees. A ratio of 1.0 would mean equal quantities of each; above 1.0 means more arabica than robusta is deliverable.

The ratio falling means the deliverable pool is tilting toward robusta — either because robusta is building, or arabica is draining, or both. The ratio alone does not say which, which is why the page also shows which leg moved it. Read it beside the absolute levels, never on its own.

Robusta as a percentage of all certified coffee across both contracts, in bags-equivalent. The same information as the ratio above, expressed the other way round — the two are locked together exactly, since share equals one divided by one plus the ratio.

Rising means robusta is taking a larger part of the deliverable pool. Because it is the ratio restated, it never tells you anything the ratio did not; it is carried because a rising number reads more naturally as 'robusta gaining' than a falling ratio does.

The gap between the two coffees' percentage moves over the same window, stated in percentage points. If robusta stocks rose 3% while arabica fell 4%, this reads +7. It measures how far the two legs pulled apart, not how far either one travelled.

Positive means robusta outperformed arabica over the window, whichever direction they both went. It is a difference of two percentage changes, not a ratio of them — a ratio flips sign or runs away exactly in the lopsided cases that matter most, which is why the report uses a difference here and a ratio for the levels above.

The percentage of all certified arabica held in whichever single warehouse location holds the most of it. A concentration measure: 100% would mean every certified bag sits in one port, a low number that the stock is spread widely.

A high or rising share means the deliverable pool is bunching into fewer places. That matters for anyone who has to physically take or make delivery — the same tonnage is easier to reach spread across many ports than piled in one. It says nothing about whether total stock is high or low.

The same concentration measure for robusta: the percentage of certified robusta sitting in its single largest warehouse location.

Read it alongside the arabica row, but remember the two describe different warehouse networks — robusta's delivery ports are European, arabica's mostly American. Where both coffees concentrate in the same port, the two legs are competing for the same physical space.

A look-back length. On this page: 1w is 5 trading days, 2w is 10, 1m is 21, 3m is 63, and 12m is 252. The right-hand columns show how far each row moved across each one.

Most rows move in percent; the divergence row moves in percentage points. The band above each group of rows states which.

A reading of certified stock alone — coffee physically sitting in exchange warehouses — as opposed to futures prices, which this report does not yet carry. It asks one question: is the deliverable pool tilting toward arabica or toward robusta, and is that tilt new?

The one-word reading at the top summarises how the time spent in a formation has changed across three successive windows, and the line beneath it gives the basis. It describes the record so far. It is not a forecast.

The direction each coffee's certified stock moved over the same five trading days, named as a pair. 'A down / R up' means arabica stock fell while robusta stock rose. There are four: both up, both down, and the two opposing pairs.

The opposing formations are the sharpest form of what this report watches for: the two deliverable pools going separate ways. In a both-up or both-down formation the ratio can still move, but only because one leg moved further than the other.

The percentage of trading days, within a window, on which the market sat in a given formation.

It describes how the period was spent, not how likely anything is. Neighbouring days share four of their five days of history, so they are far from independent observations, and no confidence range attaches to the figure.

How many consecutive weeks a formation held before changing, counted in non-overlapping weekly blocks so that no day is counted twice.

A run already under way when the archive begins, or still open at the latest close, is cut short by the edge of the data. The page says when either applies.

Each coffee's change in certified stock over the stated window, in bags, and its share of the combined movement. Movement is counted without sign, so a fall of 3,000 bags and a rise of 3,000 bags weigh the same.

The ratio cannot tell you which coffee is moving it. This can. A ratio that falls because robusta is arriving describes a different market from one that falls because arabica is leaving.

Where the current figure sits between the lowest and highest values in the archive, from 0% at the low to 100% at the high.

On a series that has trended one way, this pins near 0% or 100% and stays there. It is not a sign that a reversal is due, which is why it is only ever shown beside the level itself and the change across the whole archive.

The change in each coffee's certified stock over roughly the last month, broken down by the warehouse port holding it and, for arabica, by the country the coffee came from. Net change adds rises and falls together; gross change counts every movement without sign.

Robusta stock can be broken down by port only. The exchange report carries no country of origin for robusta, so there is no robusta table by origin.

The same four formations as the physical read, measured inside a single warehouse port that holds both coffees.

Only ports licensed for both contracts can carry a formation. Arabica's delivery network is mostly American and robusta's is European, so the shared list is short — that is a property of the two contracts, not a gap in the data.

A single number for how tightly a coffee's certified stock is packed into a few warehouses. Take each port's share of that coffee's total as a percentage, square each one, and add them up. Everything in one warehouse scores 10,000; the same tonnage spread evenly over ten ports scores 1,000. The daily figure for each coffee is published in this issue's JSON download. It is the Herfindahl-Hirschman Index, borrowed from competition analysis, where it measures how few firms hold a market.

Higher means more of the coffee is in fewer sheds — which matters to anyone who has to physically take or make delivery, and says nothing about whether total stock is high or low. The thresholds used in competition analysis are calibrated to firms in a market and mean nothing here.

The lowest HHI a coffee could score given how many ports hold it: 10,000 divided by the number of active ports. The excess is how far the actual figure sits above that floor.

Compare arabica with robusta on the excess, never on the raw HHI. Stock held in three ports cannot score below 3,333 however evenly it is spread; stock held in six cannot score below 1,667.

The International Coffee Organization's monthly certified-stock figures for New York (arabica) and London (robusta), set beside the ICE warehouse archive this report is built from. Both a month-end reading and a month average are shown, because the ICO does not say which one it publishes.

Agreement is a check that the archive is being read correctly, not a second source of news. Where the two disagree, the ICE figure is the one every other number on this page is made of.