The internet loves the “Dutch = dairy = tallest people” story. Even the BBC once framed the Netherlands as “a nation of tall cheese-eaters,” musing about whether milk and cheese helped push the Dutch to the top of the height charts - while also noting other contenders like genetics, better medical care, and even sexual selection favoring taller men.
I combined cohort-level height data with per-capita calorie/macronutrient supply (ages 0–18), then tested nutrition alone vs. nutrition + a country-mapped height polygenic score (PGS).
What I looked at
Outcome: Average male height by birth cohort (measured around age 18).
Nutrition exposures (ages 0–18):
total calories, total protein (animal + plant), and the protein-to-carbohydrate ratio (logged).
I decomposed each exposure into within-country changes over time and between-country long-run means.Genetics proxy: Country-mapped polygenic score (PGS) labels matched to a compiled height PGS file. After cleaning & matching, the PGS covered 46 countries in the analysis set.
Model: Linear mixed effects with a random intercept for country.
Nutrition-only model
Nutrition + PGS model
What the nutrition histories actually show
Key fact: The Dutch do not eat more protein or more calories than other rich countries. In our series, the U.S. is higher than the Netherlands on both total calories and total protein, and the Netherlands tracks roughly level with Denmark and Spain. The Spanish also have higher protein to carbs ratio than the Dutch (Fig. 3). So the “they’re taller because they eat more” claim doesn’t hold.
Figure 1. Total calories over time: Netherlands vs. Denmark, United States, Spain, Japan, India
Figure 2. Total protein calories over time
Figure 3. Protein-to-carbohydrate ratio (log)
A fair pushback is that the earlier charts track total protein, not dairy. Because dairy is counted under animal protein, I repeated the comparison using animal protein alone - and the Netherlands still doesn’t stand out among other wealthy countries (see the next chart).
Figure 4. Animal protein intake
These trajectories show the Netherlands and other Western countries rising early in both total calories and protein share, with Japan and India catching up more slowly. If dairy were the key, you’d expect the Dutch advantage to vanish once we control for richer protein profiles.
A cross-country snapshot: animal protein share vs height
To step back from year-to-year noise, I also collapse the data to one dot per country: the average male height vs the average share of animal protein in total protein (ages 0–18). The scatter below shows that while there’s some relationship, a high animal-protein share isn’t enough to explain who ends up tallest - countries with similar (or even higher) shares don’t uniformly top the height rankings.
What the models say
Nutrition-only model
When I regress height on within- and between-country nutrition (calories and % protein), the model explains a lot of the variation, and the country random intercepts capture the leftover, country-specific height advantage/disadvantage that nutrition can’t explain.
What the nutrition-only model shows
Data & model. 7,884 birth-cohort observations from 146 countries; mixed model with a country random intercept predicting male height from calories and % animal protein (both decomposed into within-country year-to-year changes and between-country long-run means).
Within-country effects (cohort shifts).
Calories: +0.37 cm per +1 SD (p < 0.001)
% animal protein: +0.60 cm per +1 SD (p < 0.001)
Between-country effects (long-run averages).
Calories (mean): +2.06 cm per +1 SD (p ≈ 2.9e-9)
% animal protein (mean): +1.42 cm per +1 SD (p ≈ 2.4e-5)
Fit & what’s left.
The model leaves a sizable country-specific residual (random-intercept SD ≈ 3.04 cm; variance 9.264), meaning some countries are ~3 cm taller/shorter than nutrition alone would predict. Cohort-level residual SD ≈ 0.92 cm.
Takeaway: richer calorie environments and a higher share of animal protein are strongly associated with taller cohorts and taller countries on average - but nutrition doesn’t fully explain cross-country differences. That remaining country effect is what we probe next when we add genetics.
Adding genetics (Polygenic Scores)
When I include the PGS, the picture becomes much clearer::
Figure 5. Country random intercepts: before vs. after PGS
Interpretation: Points below the diagonal indicate countries whose “mystery height bonus” drops once genetics is included. The Dutch point sits below that line - nutrition alone didn’t fully account for their height; adding PGS moves us much closer.
This is because the Dutch have the highest Height polygenic score in the world, as shown in the lollipop chart below:
Figure 6. Height polygenic scores
Here’s another way to think about it: don’t look at percent reduction - look at the absolute change.
Definition.
abs_change = |ri_nut| − |ri_pgs|measures how much of a country’s previously unexplained height gap (from the nutrition-only model) gets soaked up when we add genetics (PGS).Interpretation.
Positive (bigger is better): more of that country’s nutrition-residual lines up with the genetics proxy. After controlling for calories and protein (or % animal protein), PGS explains more of the leftover gap.
Zero: adding PGS didn’t move the unexplained country effect.
Negative: the unexplained effect actually got larger once PGS was included.
What this isn’t: a verdict that “nutrition is worse for growth than genetic potential.” It doesn’t rank nutrition vs. genetics. It shows that, given the nutrition variables in the model, part of the country-level leftover is captured by PGS.
When I rank countries by the size of the absolute change, the Netherlands sits at the top: a 2.42 cm reduction in the unexplained gap after adding PGS.
Below is a lollipop chart of the shrinkage by country (sorted high → low). Positive bars mean the unexplained country effect shrinks with PGS; negative bars mean it grows.
Figure 7.
So, is it dairy?
Nutrition helps - richer calorie and protein environments support growth - but Dutch height isn’t explained by eating more calories or (animal) protein than other developed countries (they don’t). Once you control for nutrition carefully, a genetic component explains a large share of the remaining Dutch advantage. The tidy “it’s all dairy” narrative doesn’t survive contact with the data.
japans calorie intake trajectory looks interesting; has something happened at the start of the 1990's to which the stagnation can be traced back to?