Recipes

    The Raw-Milk Wars

    Every other page in this wing is settled history. This one is a fight that is still going on, in courtrooms and state legislatures, this year — including in Michigan, where I live.

    I also have a stake in it, because my first cheese was made with raw milk from the farm down the road. So I want to be careful here. What follows is not a case for raw milk or against it. It’s an attempt to work out what the rules actually say, where they came from, and whether they do what they’re advertised to do — because on that last question the answer turns out to be genuinely uncomfortable for both sides.

    Why pasteurisation happened

    Start by giving the other side its strongest case, because it’s very strong.

    Milk is close to an ideal growth medium, it comes out of an animal, and it used to arrive in cities after hours in the heat. Urban milk supplies in the nineteenth century killed a great many people, most of them children — tuberculosis, typhoid, diphtheria, scarlet fever, brucellosis, and ordinary catastrophic diarrhoea. Pasteurisation is one of the most effective public-health interventions in the history of food, and anybody arguing about raw milk who skips past that is not arguing honestly.

    This section has already met the evidence in a form you can hold. The oldest cheese anyone has ever analysed — the jar from the tomb of Ptahmes, on the ancient world page — carried Brucella melitensis, the organism that causes brucellosis, which passes to people almost entirely through unpasteurised dairy. The hazard is not theoretical and it is not new. It is 3,200 years old and we have the sample.

    Hold on to Brucella. It comes back.

    What the rules actually are

    Two federal pieces, and they fit together more precisely than most of the shouting suggests.

    The interstate rule. In 1987 the FDA issued the final rule implementing a court order in Public Citizen v. Heckler — so it arrived through litigation rather than agency initiative. 21 CFR 1240.61 provides that no one may deliver into interstate commerce any milk or milk product in final package form for direct human consumption that has not been pasteurised.

    And then the exception. That same regulation carves out products covered by “alternative procedures to pasteurization … provided for by regulation, such as in part 133 of this chapter for curing of certain cheese varieties.”

    Part 133 is the cheese standards, and the alternative procedure is the famous one: cure the cheese for at least 60 days at a temperature above 35 °F. Promulgated in 1950.

    So the 60-day rule is not a separate raw-milk policy. It is the exception written into the pasteurisation mandate — the one door left open. Which makes the question of whether it works rather important.

    🔴 The 60-day rule was designed to kill brucellosis

    Here is the finding that made this page worth writing, and it closes a loop I didn’t expect.

    The 60-day threshold rests on decades-old research showing that Brucella abortus is eliminated in cheddar over that period. That was the pathogen of concern when the number was chosen, and against that pathogen the number was reasonable.

    The rule was calibrated against the same organism we found in the oldest cheese in the world. Three thousand years apart, Brucella is the thing the regulation is still shaped around.

    The problem is that Brucella is no longer the main threat in a country with a controlled dairy herd, and the pathogens that are the main threat behave differently.

    And it does not do what it says

    This is where it gets uncomfortable, and I want to state it precisely rather than rhetorically, because the evidence cuts against both camps.

    • 1960s challenge studies found Salmonella surviving past 60 days. Salmonella Typhi survived in stirred-curd granular cheddar for 150 to 180 days at refrigeration temperatures — two to three times the legal threshold.
    • A 2010 study in the Journal of Food Protection documented E. coli O157:H7 surviving in aged gouda and stirred-curd cheddar, with the authors concluding that a 60-day requirement founded on the Brucella research “is insufficient to completely eliminate levels of viable E. coli O157:H7.”
    • On soft cheeses the rule fails on its own terms. A study whose title says it outright: the 60-day requirement does not ensure safety of surface-mold-ripened soft cheeses — made from raw or pasteurised milk — when Listeria monocytogenes is introduced as a post-processing contaminant.

    That last one deserves reading twice, because it is the single most misunderstood fact in this whole argument. Post-processing means the Listeria got onto the cheese after the make — from a surface, a cloth, a hand, a cave. If the contamination happens after the milk stage, then pasteurising the milk would not have prevented it. The cheese is equally at risk either way.

    So the raw-versus-pasteurised axis, which is the entire public argument, is often the wrong axis. What actually predicts risk in a finished cheese is moisture, pH, salt, and how cleanly it was handled after it left the vat. A high-moisture, low-acid, surface-ripened cheese is the risky category regardless of what happened to the milk. A hard, low-moisture, well-acidified wheel aged a year is a hostile environment for pathogens whichever milk it started from.

    None of which means raw milk carries no additional risk. It plainly does at the milk stage, which is exactly why soft raw-milk cheeses are the genuinely higher-risk product and why people who are pregnant, elderly, very young or immunocompromised are advised to avoid soft cheeses generally. The honest summary is narrower and less satisfying than either slogan:

    Sixty days is a 1950 proxy for a pathogen that is no longer the main problem, it does not reliably control the pathogens that are, and the raw/pasteurised binary is a poor predictor of the risk in a finished cheese compared with the cheese’s own moisture, acidity and handling.

    The rule survives because it is simple, administrable, and better than nothing — not because it is accurate.

    Europe took the opposite bet

    The instructive contrast is that another sophisticated regulatory system looked at the same food and went the other way — not by tolerating raw milk, but by requiring it.

    Roquefort — the first French cheese to receive protected status, in 1925 — is defined as being made from raw whole sheep’s milk. Pasteurise it and it is not Roquefort. Several other PDO specifications work the same way.

    And look at the geography of the modern rule, because it is the 1411 charter again: as on the monasteries page, the milk may be drawn from a wide surrounding area — around a 60-mile radius — but the cheese can only be aged in the village itself to qualify. Six hundred years on, the protected asset is still the cave.

    Two systems, the same food, opposite conclusions: one treats raw milk as a hazard to be permitted only under a time rule, the other treats it as the defining characteristic to be legally mandated. They are not reading different science so much as weighting different things — one optimising for the worst plausible outcome across a huge industrial supply, the other for the character of a specific product made by a specific method.

    Michigan, which is where I actually live

    Michigan is at the strict end, and the mechanism is worth knowing precisely.

    Selling raw milk is illegal in Michigan — no grocery stores, no retail, no direct farm-to-consumer sale.

    But a 2012 state task-force policy allows you to consume unpasteurised milk from a cow you own, either outright or through a herdshare. In a herdshare you pay a farmer to board, care for and milk the animal, and as a shareholder you then obtain — but do not purchase — the milk. The distinction is the whole legal architecture, and it comes with conditions: a signed and dated written contract, established communication between farmer and shareholder, and milk from a single farm, not comingled.

    Two things follow that I find genuinely interesting. Herdshares are not licensed, so the state has no idea how many exist — MDARD says as much. And the arrangement is under active review: Michigan lawmakers were considering changes to the raw-milk laws as recently as March 2026, so anything in this paragraph may date.

    Meanwhile the cheese question is separate and federal: a raw-milk cheese aged 60 days is legal, here and everywhere in the US, under the part 133 exception above.

    Where I land

    I’m not going to pretend to a tidy position, because I don’t have one and the evidence doesn’t support one.

    What I’m confident of: the 60-day rule is a bad instrument — miscalibrated to the wrong organism, demonstrably insufficient against the right ones, and blind to the distinction that actually predicts risk. Replacing it with something built on moisture, pH and validated process control would be better public health and better for good cheese, which is a rare combination.

    What I won’t say: that the risk is invented. It isn’t, and the people insisting it is are usually arguing from ideology. Soft raw-milk cheese carries real hazard, the vulnerable groups are genuinely vulnerable, and Brucella has been in our cheese for three thousand years.

    And for my own bench: the cheese I make is for my family, from milk I can trace, aged in a cave I control, and I’d make different choices for a soft bloomy rind than for a hard wheel I’m going to keep a year. That distinction — by cheese type, not by slogan — is the one this page argues the regulation itself should be making.

    Up next: today — where the craft actually stands, and what I think happens next. The last page of the history wing.

    Sources

    Comments (0)

    1. Loading…

    Comments are held for moderation — nothing appears until approved.