Carbs vs fat in a calorie deficit: which should you actually cut?

Every few years the fitness industry flips which macro is the enemy. In the 1990s it was fat — low-fat cookies, low-fat yogurt, low-fat everything. In the 2010s the pendulum swung to carbs — keto, low-carb, paleo. In the mid-2020s it's back to moderate everything and "focus on whole foods."

The underlying research, though, has been boringly consistent for 30 years: calorie balance determines fat loss; macro composition determines how you feel while losing it. Here's the honest version — what the studies actually show, and how to choose the cut that fits you.

The short answer

For pure fat loss at a matched calorie deficit, low-carb and low-fat produce essentially identical results over 6–12 months. The difference between groups in randomised controlled trials is typically 0.5–1.5 kg over a full year — statistically there, practically trivial.

What isn't trivial is which one you can sustain. Adherence — not macro magic — is the variable that separates the two.

The "low-carb wins" advantage, where it exists, is almost entirely a water-weight artifact in the first 2 weeks. Each gram of glycogen in your body is bound to ~3g of water. Drop carbs, drop glycogen, drop 1–2 kg of water. Looks great on the scale. Doesn't move fat mass faster than any other approach.

Why the industry keeps flipping

Short-term studies (2–8 weeks) tend to favor low-carb because of the water-weight trick and because novel dietary patterns have a honeymoon effect on adherence. Long-term studies (6–24 months) consistently show convergence.

So researchers keep publishing "low-carb is better" at 8 weeks, then "no difference" at 12 months, and the media picks whichever headline they need that season.

When low-carb actually wins

Not for everyone, but genuinely for some:

1. Insulin resistance / pre-diabetes

If fasting insulin is elevated, or HbA1c is in the pre-diabetic range (5.7–6.4%), lower-carb eating modestly outperforms low-fat for fat loss. Research puts the edge at about 10–15% better adherence and slightly better metabolic markers over 6 months. Not a miracle, but a real signal.

If you've never had bloodwork, you don't know. Most people tracking weight loss aren't insulin resistant — they just assume they are because low-carb is marketed as a fix for it.

2. Crave sweet / carby foods constantly

If "just one slice of bread" regularly becomes half a loaf, going low-carb eliminates the trigger food category entirely. This is a behavioral advantage, not a metabolic one. Some people find it easier to cut carbs to 100g/day than to eat them "in moderation" all day.

3. Training is recreational, not performance-based

If you lift 3× a week for general health and cardio is walking or easy cycling, low-carb works fine. Your performance ceiling isn't where it needs to be for glycogen to matter much.

When low-fat (relatively) wins

1. Endurance athletes or high-volume training

Running 40+ km/week, cycling 100+ km/week, or training for an event? Glycogen fuels that work. Low-carb athletes famously lose intensity — they can still do long slow efforts, but top-end performance degrades within a few weeks of going below 100g carbs/day.

A moderately lower-fat approach (25–30% of calories) preserves training fuel and still creates a significant deficit.

2. You love bread, pasta, rice, and fruit more than you love cheese, butter, and fatty meat

Adherence is the variable. If carb foods are what make eating satisfying for you, a low-fat deficit is usually easier to sustain. Don't fight biology and your preferences.

3. Cardiovascular risk factors

If your doctor has flagged elevated LDL, lower-saturated-fat eating patterns (which usually end up lower-fat overall) have 40 years of cardiovascular research behind them. A strict keto diet high in saturated fat isn't the right call for someone with a family history of heart disease.

The two floors nobody talks about

You can't cut either macro to zero. Each has a minimum:

Fat floor: ~0.5 g/kg bodyweight

For a 75 kg person, that's ~38g fat/day. Below that:

• Hormones suffer — testosterone and estrogen drop measurably within 3–4 weeks
• Fat-soluble vitamins (A, D, E, K) stop absorbing properly
• Satiety collapses — fat is the slowest-emptying macro; without it meals don't hold

Most "low-fat" diets sit at 0.6–0.8 g/kg, which is fine. The danger zone is 20g/day fat and below — common in extreme low-fat attempts. Avoid.

Carb floor: ~100g/day (for most adults)

You can go lower — ketogenic diets run 20–50g carbs — but below 100g most people experience:

• Training performance drops 10–20% within 2 weeks
• Brain fog / "keto flu" for the first 7–14 days
• Sleep disruption (carbs at dinner modestly improve sleep)

Below 50g, you're in a metabolically distinct state (ketosis) with its own rules. That's a choice, not a default. For most people chasing fat loss, 100–200g carbs/day is the practical floor.

The real lever: protein (neither carb nor fat)

The macro that matters most for how a deficit feels isn't carbs or fat. It's protein. In every study where protein was held high (≥1.6 g/kg), the carb/fat split mattered much less. When protein was low (under 1.2 g/kg), both groups felt miserable.

Prioritize protein first, fill the rest with carbs or fat based on your preferences and training. Full protein-target math is here if you haven't set one.

A practical decision tree

Use this in 60 seconds:

1. Are you a competitive or volume-heavy endurance athlete? → Keep carbs higher (40–50% of cals). Cut fat modestly.

2. Do you have insulin resistance, pre-diabetes, or strong cravings for carbs? → Try moderate low-carb (100–150g carbs/day). Easier for most people with those traits.

3. Do you love carb-heavy foods (rice, pasta, bread, fruit) and training is recreational? → Go lower-fat (25–30% of calories). Still plenty of room for the foods you like.

4. None of the above particularly applies? → Balanced Mediterranean-style: 30% protein, 30% fat, 40% carb. Easiest to sustain for most people, strongest long-term evidence.

None of these are "optimal" in some absolute sense. They're sustainable — which is the only metric that matters over 6 months.

What about keto specifically?

Keto (under 50g carbs/day, inducing ketosis) has a specific, narrow use case:

• People with seizure disorders (it's a recognized medical intervention)
• People who genuinely cannot moderate carbs behaviorally and find abstinence easier than limits
• A short 4–8 week "cut" phase where the appetite-suppression is a known benefit

It is not metabolically superior for fat loss in most studies. It has high dropout rates (often 50%+ at 12 months). And the long-term cardiovascular impact of a diet high in saturated fat from meat and dairy is still contested.

If you like keto, it works. If you're wondering whether you need to do keto to lose weight — you don't.

The boring, evidence-backed answer

1. Set a reasonable calorie deficit — 500 kcal below TDEE
2. Hit 1.6 g/kg protein
3. For carbs and fat, use whatever ratio you can sustain for 12 weeks
4. Don't go below 0.5 g/kg fat or 100g carbs (unless you're deliberately doing keto)
5. Ignore the macro debate; fix the under-logging and weekend gap instead — those matter 5× more than whether your rice was replaced with cauliflower

Macro composition is a preference, not a magic lever. Pick the one you can eat honestly for three months. That beats the "optimal" one you'll quit in six weeks.

Pairs well with: how much protein you actually need (the macro that does matter), and what a sustainable daily deficit looks like.