Why We Use Three Different Fats

Palm shortening, ghee and neutral oil each do a job the other two can’t — and none of them were chosen for a health claim

People assume a bakery picks a fat the way you’d pick a paint colour — one decision, applied everywhere. It doesn’t work like that. Fat is the most structurally demanding ingredient in the building, and asking one of them to do every job is how you end up with a croissant that behaves like a scone.

We use three: a palm-based shortening, ghee, and a neutral oil. We’re consolidating the shortening side onto a single supplier, which I’ll explain in a companion piece, because the palm question deserves its own article rather than a paragraph. This one is about the craft: what these fats actually do, and why the answer has almost nothing to do with the health claims printed on them.

“Shortening” is a verb, not a marketing word

Start with the thing nobody explains. The word shortening describes an action. As the standard food-science reference on lipid shortenings puts it, the term “refers to the ability of a fat to lubricate, weaken, or shorten the structure of food components.”

Here’s the mechanism, and it’s satisfying once you see it. Flour plus water builds a continuous network of gluten and starch. Left alone, that network is strong — which in a bread is the point, and in a shortbread is a disaster. Without fat, “gluten and starch particles adhere to each other and give the sensation of hardness and toughness when chewed. However, if shortening is present, the fat breaks the continuity of the protein and starch structure.”

The fat gets in the way. That’s the whole trick. Tenderness is interruption.

Why solid fat, specifically

This is where it stops being about fat and starts being about crystals.

A solid fat isn’t uniformly solid. It’s a suspension — fat crystals held in liquid oil — and the proportion of crystal to liquid shifts as the temperature changes. A general-purpose bakery shortening runs somewhere around 15–30% solid across normal working temperatures. Roll-in shortening for laminated pastry needs 40% or more when cold, and rather less when warm.

That curve is the entire specification. A fat isn’t chosen for how hard it is; it’s chosen for how its hardness changes as it warms. Which is why the same fat cannot laminate a croissant, cream a cake batter, and hold up a buttercream. Those are three different curves.

Then there’s crystal shape, and this is my favourite piece of food science because it’s so counterintuitive. Fats are polymorphic — the same molecules can pack in different arrangements depending on how they’re crystallised. Three forms matter, in increasing order of stability: alpha, beta-prime, and beta.

Beta-prime is the one bakers want. Its crystals are tiny — around a micron — and needle-shaped, and that geometry is exceptionally good at trapping and holding air. Beta crystals, the more stable form, grow into “large plates of solid fat” that are “much less effective in entrapping dispersed air.” A cake shortening that drifts from beta-prime to beta under fluctuating storage temperatures loses baking performance. Same fat. Same fatty acids. Worse cake.

The clearest demonstration of this comes from lard. Natural lard performs badly in cakes because of where palmitic acid sits on the triglyceride molecule. Rearrange those fatty acids — same atoms, same saturation, different arrangement — and the fat converts from beta to beta-prime, which “allows better incorporation of air into cake batters yielding increased cake volume.”

Nothing was added. Nothing was removed. The molecules were reshuffled, and the cake got taller.

I find that genuinely clarifying, because it cuts against how we’re all trained to read food. A fat’s performance is a question of physical structure, not just its saturated-versus-unsaturated ratio. The label tells you almost nothing about what the fat will actually do.

What ghee is doing here — and what it isn’t

Ghee is in our kitchen for flavour and for physics, and I want to be straight about that, because ghee has become a wellness product and the evidence doesn’t support the halo.

The functional case is real. Ghee is milk fat with the water and milk solids taken out — at least 99.6% milkfat under the international standard, against butter’s 80% fat and up to 16% water. That purity matters in a formula. Butter smuggles water into your recipe; ghee doesn’t. Because the milk proteins and sugars that scorch have been removed, it also tolerates heat better than butter does. And the browning those solids undergo during manufacture is exactly where ghee’s nutty, slightly caramel flavour comes from — a flavour you cannot get from butter without burning it.

One genuinely useful fact: ghee is essentially lactose-free. Butter carries around 685 mg of lactose per 100 g; measured ghee comes in under 3 mg, and often below the detection limit. That’s a roughly 250-fold reduction, and it’s why butter oil is permitted in some clinical diets where butter isn’t.

But — and this matters enormously — that does not make ghee safe for a milk allergy. The FDA’s position is that ghee commonly contains residual milk protein and must carry a milk allergen declaration. Lactose is a sugar; casein is a protein. Removing one does not reliably remove the other. Intolerance and allergy are different conditions, and anyone conflating them in a bakery is going to hurt someone.

Now the part ghee marketing skips. In the best available human trial — a randomised crossover study in which healthy adults ate about 30 g of ghee a day for four weeks — ghee raised apolipoprotein B and non-HDL cholesterol compared with olive oil. The one meta-analysis specific to ghee is inconclusive, drawn entirely from studies on the Indian subcontinent, and published in a low-impact journal. There is no good evidence that ghee is a heart-health food. It is roughly 70% saturated fat, and clarifying it doesn’t change that — removing water and milk solids concentrates the fat, it doesn’t restructure it.

So we use ghee because it tastes like something no other fat tastes like, and because its purity behaves predictably. Not because it’s virtuous. If we told you otherwise we’d be doing exactly what we criticise other people for.

What the neutral oil is for

“Neutral oil” is kitchen shorthand, not a technical category — there’s no regulatory definition of it. What it describes is the outcome of refining: deodorisation strips out the volatile compounds that would otherwise carry the seed’s own flavour. Worth being honest that this is a trade-off, not a free lunch. That same refining removes some tocopherols and phytosterols along with the flavour. It’s also what makes the oil stable enough to be worth using.

Its job in our kitchen is cake. Here the conventional wisdom that solid fat always wins is simply wrong, and it’s been measured: in a study comparing high-ratio layer cakes made with plastic shortening, liquid shortening, plain oil, and oil plus emulsifiers, the liquid systems performed best on fresh cake characteristics and firmness over time. Where a fat’s job is aeration and that job can be shifted onto emulsifiers and chemical leavening, liquid oil competes and often wins.

Where liquid oil cannot compete is anywhere the fat’s job is to build and hold a crystal network — lamination, pie flakiness, the body of an icing, the structure of a cookie that has to survive a box. That distinction explains something every home baker has noticed without being told why: oil cakes stay soft for days, and oil croissants don’t exist.

A detour, because someone always asks

Since our neutral oil is a seed oil, let me deal with the claim that seed oils drive inflammation, because it’s everywhere and it doesn’t hold up.

The direct test has been done. A systematic review of fifteen randomised controlled trials in healthy adults, measuring a wide range of inflammatory markers — C-reactive protein, fibrinogen, TNF-α and others — found no evidence that adding linoleic acid to the diet raises them. None of the studies reported significant findings. Separately, a pooled analysis of thirty cohorts across thirteen countries — 68,659 people, over 15,000 cardiovascular events — found that people with more linoleic acid in their blood had lower cardiovascular risk, not higher. And the American Heart Association notes that average US omega-6 intake sits at about 6.7% of calories, which is inside the 5–10% range it recommends, not wildly beyond it.

But I’ll give the other side its due, because overselling is the same sin in reverse. The Cochrane review on omega-6 fats found high-certainty evidence that they lower total cholesterol — and low-certainty, non-significant results for whether eating more of them actually prevents heart attacks or deaths. “Not harmful” and “actively protective” are different claims. The first is well supported. The second isn’t.

We don’t use neutral oil because it’s a health food either. We use it because it makes a better cake.

The through-line

Three fats, three jobs, none of them chosen from a label.

The shortening builds structure and holds air, because of a crystal form most people have never heard of. The ghee brings a flavour and a purity nothing else brings. The oil keeps a cake tender on day three.

That’s the actual reasoning. It’s less tidy than “we use the healthy fat,” and I think it’s more useful — because once you know that fat performance is about crystal structure and melting curves, you can never again be sold a fat purely on the adjective printed above it.

Which brings us to the fat we use most of, and the hardest question we’ve had to answer about any ingredient in this bakery. That’s the next piece.

Where these numbers come from

• Definition and mechanism of shortening; plasticity and solid fat content; crystal polymorphism (alpha, beta-prime, beta), 1 μm beta-prime crystal size, and the loss of cake performance on conversion to beta: Ghotra BS, Dyal SD, Narine SS, “Lipid shortenings: a review,” Food Research International 35:1015–1048 (2002) — talcottlab.tamu.edu
• Solid fat content ranges for all-purpose and roll-in shortenings: Oklahoma State University Extension / FAPC, “Shortenings,” Fact Sheet FAPC-212 — extension.okstate.edu
• The lard randomisation example (beta to beta-prime, increased cake volume): American Oil Chemists’ Society, “Trans Fat Replacements in Foods” — aocs.org
• Ghee composition minimum of 99.6% milkfat; butter at ≥80% fat and ≤16% water: Codex Alimentarius (FAO/WHO), Standard for Milkfat Products CXS 280-1973 and Standard for Butter CXS 279-1971 — fao.org
• Lactose content of butter (685 mg/100 g) versus ghee (<3 mg/100 g): “The lactose and galactose content of milk fats and suitability for galactosaemia,” Molecular Genetics and Metabolism Reports (2015), PMID 28649541
• Ghee commonly contains residual milk protein and requires milk allergen declaration: US Food and Drug Administration, Questions and Answers Regarding Food Allergens, Q. B.7 — fda.gov
• Ghee versus olive oil randomised crossover trial (raised apolipoprotein B and non-HDL cholesterol): Hosseinabadi SM, Nasrollahzadeh J, British Journal of Nutrition 128(9):1720–1729 (2022)
• Milk fat approximately 70% saturated: Lindmark Månsson H, “Fatty acids in bovine milk fat,” Food & Nutrition Research 52 (2008)
• Refining removes tocopherols and phytosterols along with flavour volatiles: Oklahoma State University Extension, “Canola Oil Properties” — extension.okstate.edu
• Liquid shortening and oil-plus-emulsifier systems outperforming plastic shortening in high-ratio layer cakes: Zhou J, Faubion JM, Walker CE, LWT — Food Science and Technology 44(8):1802–1808 (2011)
• Fifteen randomised controlled trials finding no effect of dietary linoleic acid on inflammatory markers: Johnson GH, Fritsche K, Journal of the Academy of Nutrition and Dietetics 112(7):1029–41 (2012) — pubmed.ncbi.nlm.nih.gov
• Pooled analysis of 30 cohorts, 68,659 participants, higher circulating linoleic acid associated with lower cardiovascular risk: Marklund M et al., Circulation 139:2422–2436 (2019)
• US omega-6 intake of ~6.7% of energy against a recommended 5–10%: American Heart Association Science Advisory, “Omega-6 Fatty Acids and Risk for Cardiovascular Disease,” Circulation (2009) — professional.heart.org
• High-certainty evidence that omega-6 lowers total cholesterol; low-certainty evidence on mortality and cardiovascular events: Hooper L et al., Cochrane Database of Systematic Reviews (2018), “Omega-6 fats for the primary and secondary prevention of cardiovascular disease”

Every figure above comes from government or intergovernmental standards, peer-reviewed research, or university food-science publications. Where a popular claim about ghee or seed oils is supported only by weaker sources, we’ve said so rather than repeating it.