Reading concentration ratios in agribusiness: CR4, the HHI and what they prove

Every so often a number goes round: four companies process most of America's beef, four traders handle most of the world's grain, a few firms sell most of the commercial seed. The figures are usually in the right neighborhood. What they mean is another matter, and it is not a question the figures can settle on their own.

This is a field guide to the two measures behind most of those claims, the four-firm concentration ratio and the Herfindahl-Hirschman Index, with examples small enough to check on the back of an envelope. After that comes the harder part: what a high reading does and does not tell you about prices and power, and where the underlying data come from.

CR4: easy to compute, easy to over-read

The four-firm concentration ratio, written CR4, is the combined market share of the four largest sellers. If the top four hold 30%, 25%, 20% and 10%, the CR4 is 85. That is the whole calculation.

Its strength is that anyone can work it out and explain it in a sentence. Its weakness is that it throws information away. It ignores how share is split among the top four, and it ignores everyone below them. A market where four firms hold 25% each and a market where one firm holds 70% and three rivals hold 10% apiece both score a perfect 100. Nobody who has tried to sell into the second market would call them equivalent.

Statistical agencies publish cousins (CR8, CR20, CR50) for the same reason a photographer carries more than one lens. But CR4 is the version that ends up in headlines, partly because "the big four" is a phrase editors like.

The HHI: let the big shares shout

The Herfindahl-Hirschman Index squares every firm's market share and adds up the results. It covers every firm in the market, not just the leaders. With shares written as whole percentages, the index runs from near zero (a crowd of tiny competitors) to 10,000 (one firm with everything).

Squaring is the clever part. A firm with 10% of the market contributes 100 points. A firm with 40% contributes 1,600: four times the share, sixteen times the weight. A long tail of 1% firms barely registers, because one squared is one.

There is a handy shortcut for mergers, too. When two firms with shares a and b combine and nothing else changes, the HHI rises by exactly 2 × a × b. Two firms with 15% and 10% add 300 points when they merge. Two firms with 1% each add two.

The index matters outside economics classrooms because antitrust agencies use it as a first screen. In December 2023 the US Department of Justice and the Federal Trade Commission issued new merger guidelines. Under them, a market with an HHI above 1,800 counts as highly concentrated, and a merger that raises the index by more than 100 points in such a market is presumed to substantially lessen competition. The 2010 guidelines they replaced drew the "highly concentrated" line at 2,500 and looked for an increase of more than 200 points. Guidelines are not statute. They describe how the agencies will build a case, and courts still decide. But the change means a given deal now trips the wire sooner.

A worked example (every number here is invented)

The four markets below are hypothetical. They are not modelled on any real industry; they exist only to show how the two measures can disagree.

Three things jump out.

First, markets A and B share a CR4 of 100, yet B's HHI is more than double A's. The single giant in B does most of the work.

Second, market C has four firms holding four-fifths of sales and still sits below the 1,800 line in the 2023 guidelines. A headline would call C a "big four" market. The agencies' first screen would not call it highly concentrated. Both descriptions are arithmetically honest, which is exactly why you need to know which one you are reading.

Third, run a merger through market D. If the third and fourth firms (15% and 10%) combine, the CR4 barely moves, from 85 to 86, because the fourth slot is now filled by a 1% firm. The HHI rises by 300 points, from 2,340 to 2,640. That deal would draw a presumption of harm under both the 2023 guidelines and the older 2010 ones. The CR4 would have let it through without a second look.

The arithmetic is rarely where the fight happens, though. The fight is over the denominator: what counts as "the market". Define it as all meat sold in the United States and almost any packer looks modest. Define it as fed cattle bought within a day's haul of one plant and the same company may be one of two buyers. Most merger battles are battles over that line.

Why the meatpackers and the grain traders keep turning up

The big four in US beef

In US beef the four are Tyson Foods, JBS, Cargill and National Beef. Together they slaughter roughly 85% of the steers and heifers fed for beef. That is the figure the White House used in January 2022 when it announced $1 billion from the American Rescue Plan to expand independent meat and poultry processing. The same announcement put the four-firm share at about 70% in pork and a little over half in chicken.

Two features make beef the textbook case. One is the trend. In the late 1970s the four largest packers handled well under half of fed-cattle slaughter, and the big jump came over the following fifteen years or so, as large plants shipping boxed beef displaced older, smaller ones. The other is the direction of the power. Packers sell to retailers, but they buy from feedlots and ranchers, and it is the buying side (economists call it monopsony) that drives most of the anger in cattle country. A rancher does not care much about the national CR4. What matters is how many packers will bid on a pen of cattle within trucking distance, and in parts of the Plains the honest answer is two or three.

Ownership adds a wrinkle. JBS is Brazilian, and National Beef is majority-owned by another Brazilian meat company, Marfrig. That changes nothing in the arithmetic, but it explains why "foreign control" sometimes rides along with the concentration complaint.

ABCD in grain

The grain traders get their own acronym: ADM, Bunge, Cargill and Louis Dreyfus, or ABCD. A 2012 Oxfam report estimated their combined share of global grain trade at somewhere between 75% and 90%. A range that wide is itself a finding. It tells you how little reliable company-level data exist for world grain flows.

The traders are hard to measure for several reasons. Two of the four, Cargill and Louis Dreyfus, are privately held, so their disclosures are thinner than a listed company's. The cast also changes. China's state-owned COFCO built a global trading arm through acquisitions in the 2010s, and in June 2023 Bunge announced a merger with Viterra, the former Glencore Agriculture, a deal still before regulators in early 2024.

Above all, a trader's power is poorly captured by a share of tonnage. Owning the export terminal on a particular river, the crushing plant in a particular province or the ships under charter in a tight season gives leverage that a global percentage averages away.

What a high number proves, and what it doesn't

Concentration describes structure. It is a reason to look closely, not a verdict. Economists spent much of the mid-twentieth century assuming that structure more or less dictated behavior and results, and the decades since have gone into finding out how often that holds. The useful position sits in between.

A high CR4 or HHI does not, on its own, establish:

• That prices are higher than they would otherwise be. Big plants run cheaper per head than small ones. Whether those savings reached ranchers or shoppers, or stayed with the packer, is an empirical question that needs price and cost data.
• That firms coordinate. Concentration makes coordination easier, but proving it takes evidence of conduct. Pilgrim's Pride pleaded guilty in 2020 to a federal price-fixing charge in the broiler chicken business, and several large meat companies have paid to settle civil price-fixing suits in pork, beef and chicken without admitting liability. Those cases turned on communications and pricing patterns, not on concentration ratios.
• That the market is drawn correctly. National shares can understate local concentration, and global shares can hide control of a single export corridor.
• That contracts are fair. In fed cattle, only a shrinking minority of animals now trades through negotiated cash sales. Most move on formula or contract terms that refer back to those thin cash prices. A ratio can't see that.

What a high reading does tell you is how much depends on a few decisions. In the spring of 2020, when COVID-19 outbreaks shut several large packing plants, cattle backed up on feedlots and their prices fell while wholesale beef prices jumped. USDA examined that widening spread in a report published in July 2020. Nobody needed a collusion theory to see what a concentrated bottleneck does when part of it closes. A high reading also tells you the next merger matters more: in a market already above the guideline line, even modest combinations deserve scrutiny.

New entrants make good stories, whether a rancher-owned plant or a vertical farm in a warehouse, but they rarely move a national ratio. Worth remembering when they are offered as proof that the market is working.

Where the numbers come from

Most concentration figures trace back to a short list of sources, each with blind spots.

• The US Census Bureau's Economic Census, run for years ending in 2 and 7, publishes concentration ratios and HHIs for manufacturing industries, meat processing included. It is careful and slow. Figures appear years after the reference year.
• USDA's Agricultural Marketing Service, home of the Packers and Stockyards Division and of the mandatory livestock price reports. Beef and pork slaughter shares usually originate here.
• USDA's Economic Research Service, which publishes studies of market structure across food and farm inputs.
• Company filings and merger decisions. Annual reports give revenue but often blur product lines. European Commission merger decisions are valuable because they discuss market shares in detail, though usually in redacted bands such as "[30-40]%".
• Trade statistics from FAOSTAT and UN Comtrade, which show flows between countries, not companies. Company shares of world trade are almost always estimates assembled by consultants, academics or NGOs. Read the footnotes.

Most of these are on the links page.

Before repeating a concentration figure, ask five questions. What market, meaning which product and which geography? Share of what: sales, volume, capacity, head slaughtered? Which year? Who compiled it, and from what sources? And how were joint ventures and minority stakes treated? A figure that survives those five is usually worth quoting. One that can't answer them is decoration.

The corporate power shelf collects everything on this blog about mergers and market shares. Seed, where the same arithmetic tells an even starker story, is next on the list.