It’s not uncommon to hear people pointedly suggest that a vegan diet (that is, entirely plant-based) would be bad for the world because it would mean more crops over more land area. These people, often farmers, say that plant-based diets require vast areas of land covered in monoculture crops, the argument being that this is bad for the environment, limits biodiversity and encourages the use of artificial fertilisers and various pesticides etc. They point to vegans needing soy for tofu and milk, almonds for almond milk, avocadoes and quinoa and so on.

The trouble with this argument is, I think, that it is rather misinformed. Right now, there are very few vegans in the world, perhaps no more than 2% of the global population or less. Of course, many people are turning to plant-based diets for a variety of reasons and  even omnivores are buying plant milks and increasing their vegie intake in an effort to be healthy. Nonetheless, given the vast amount of meat eaten in the world and the extensive nature of animal farming, the real world effects of these trends are vanishingly small when compared to the impacts of everyday agriculture at the global scale.

On the other hand, vegans like to point to animal farming as having major negative impacts on the world. They note that about 70% of all the soy grown is for animal feed and that overall, perhaps as much as 50% or more of all cereal grains are grown to feed animals. On top of this, nearly 85% of all farmland is used to raise animals. 

What is the truth? Well, it’s not that easy to tease out. But at a sort of generalised, indicative level, some things are clear. The first is that the vast areas of monocropping are not the fault of plant-based diets but rather are the result of what all of us are doing. Secondly, right now plant-based diets aren’t really making much difference to anything. Not at the global scale, anyway. And thirdly, the impact of animal farming is truly extensive.

All of that said, I wondered at one simple question. IF the world went plant-based tomorrow and animal farming (and seafood eating) stopped entirely, would the world have more land under crops? The answer, it seems, is no. In fact, it most probably would have less. Eliminating animal farming would probably mean two things – first, a LOT of land available for rewilding or other uses, including the growing of crops for food and secondly, there would be less land covered by monoculture cropping. Mind you, this latter claim is not as bold as it might first seem.

How do I come to this conclusion? Read on. 

The first thing we need is a baseline. According to the “Our World in Data” website, there are about 1.65 billion hectares of crops in the world. Of this, the main crop types are:

Cereals (eg wheat and rice): 700 million hectares.
Coarse grains (eg maize, barley, sorghum etc): 350 million hectares
Oil crops (eg palm, soy, canola etc): 300 million hectares
Pulses: 100 million hectares
Tubers, fruits, vegetables: about 80 million hectares each
Nuts: about 20 million hectares

Clearly, cereals such as wheat and rice dominate (wheat alone uses somewhere around 250 million hectares globally). The next largest crop type, coarse grains, are largely a food source although in the OECD they are used primarily for animal feed (perhaps as much as 60-70% of all such grains grown in the US are for animal feeds). Oil crops produce cooking oil as well as food ingredients, biofuels and some industrial applications.

One of the interesting facts we learn is that while vegans often claim that 70-80% of all soy is grown for feed, this isn’t strictly true. In fact, almost all soy is grown for human uses. What is happening is that most soy is grown to service two markets and what a grower is interested in is the prices at market. At the end of the day, around 87% of all soy grown is crushed for oil for human use in cooking, biofuel and industrial applications. Of that 87%, about 20-30% by weight is recovered as oil and the residue, 70-80% by weight, is used as meal cake for animal feed (the latter is highly sought due to the optimal protein profile). Just 7% or so is grown directly for animal feeds and around 6% for human food. The remarkable growth in soy however is driven by the demand for animal feed (in other words, other oil crops are grown less and soy grown more, so it’s relative proportion of the total mix of oil crops has increased). Today, there is around 100 million hectares under soy globally, including quite a bit of South America.

The second thing is that of the coarse grains, some very large proportion is used or grown as feed. Corn, for example, is largely grown as feed or biofuel stock in the US. While grains such as maize and millet feature heavily in diets is poorer nations, at the global scale it seems that other uses are perhaps as important. Somewhere between 30% and 50% are used as animal feed (and perhaps another 10-15% for biofuel and other industrial applications). On this question of food, quite a bit of wheat also services the animal feed industry (for example, of domestic uses for wheat in Australia, about 60-70% is for animal feed). Globally, it is possible some 10-20% of wheat is grown for feed, however it isn’t clear just how much. Wheat IS primarily grown for food, but poor quality yields and residues are also sold to feed. I have estimated about 10% directly grown for feed.

Current vegan demand for various foods simply doesn’t register at this scale. The biggest demands are clearly human food first (ie everyone, not just vegans) and livestock feed. From this it is clear that almost all of the land under crops is due to general human demand, whether that be for food, animal feed, oils, biofuels and so on.

To ascertain the effect of removing animal feed from the equation, I followed this method. First, I calculated how much land is likely to be freed up once demand for feed is removed. I am not concerned with what happens to pasture or grazing land, just the question of land area under crops. Then, I calculated how much land would need to be devoted to crops to replace the lost food (in this, I am assuming seafood is included in the various numbers I uncovered – this may be an invalid assumption). The difference between these two results gives us the actual effect.

So, let’s look at cropland saved from the elimination of animal feed.

Some large proportion of all wheat grown is used for animal feed, however this isn’t all derived from wheat specifically grown for feed. So determining just how much might be saved by no longer needing wheat as feed is not easy. I will use 10% but could be persuaded it is more. 10%x250 million = 25 million hectares.

I have accepted that about 30% of all coarse grains grown globally are used for feed (this is somewhat debateable as I said, but most sources agree that it is somewhere around that value – those that suggest it is as high as 50% are including wheat in the calculation). If so, 30%x350 million Ha = 105 million hectares.

Oil crops are used for cooking, biofuel and industrial applications. I have estimated about 100 million hectares under soy, the balance (200 million Ha) is a mixture of other oil crops. If soy meal was no longer marketable, it is likely the market would swing in favour of other oil crops such as palm oil due to the greater yields per hectare. Is this really likely? I suspect so, but the extent to which this might happen isn’t clear. I am going to assume that about 50% of all soy crop would be replaced by palm crop.

87% of soy is used for oil and feed, assume 50% replaced by palm. This would save around 34.5 million hectares (ie about 34.5% of all soy grown). Saving = 34.5 million hectares.
7% of soy is used directly as feed, so that is an upfront saving. 7%x100 million = 7 million hectares.
6% of soy grown (about 6 million hectares) is directly for human food but this would remain so it won’t represent a saving..

Total reduction in area under soy comes to 41.5 million hectares.

Now, some might argue that by eliminating meat we would substantially reduce the demand for oil in cooking. I don’t think so. Oil is mostly used for frying, baking and as an ingredient. Baking and ingredient use should remain the same. As for frying, I see no reason to expect that the profile of food consumption would change substantially – that is, people would still eat fast foods, eat in at restaurants and enjoy home cooked meals and barbecues. The emerging market for “fake” meats suggest that to be the case (eg plant-based burger patties, pretend chicken, fish and the like). And of course, french fries would remain in huge demand. So I have allowed no offset for a reduction in oil demand.

The total reduction in land under crops comes to 171.5 million hectares. If however we use 50% as the extent of land under coarse grains used for feed, this increases to around 240 million hectares.

Total reduction = 170 million hectares (240 million hectares)

Now, we need to know how much extra plants we need to grow for human food. Here, I wasn’t certain how to proceed. However, the loss of meat primarily represents a loss of protein. This does have to be made up, so I have decided to use protein as the overwhelmingly principle replacement food source. That means we’d replace meat with high protein crops such as pulses and legumes. Now, we know that Westerners tend to vastly over-consume protein but in poorer countries the opposite is true. Taken as a global average, I think this comes out about even so the best statistic to use is the FAO’s estimation of average protein source by type. Those numbers are: 57% from plants and 43% from animals. Using this suggests we need to grow enough high protein crops to generate 43% more protein.

Also averaged globally we have:

Weight for one person: 62kg
Daily RDI for protein: .8g/kg
Global population: 7.8 billion.

An average person therefore needs 50g of protein daily, or 18.25kg per year.
Global protein demand = (43%x18.25kg) x 7.8 billion = 61 billion kg annually.
The best plant sources deliver about 20% protein by volume, so 5 x 61 billion = 305 billion kg
Crop yields average about 2500kg/Ha, so 305 billion/2500 = 122 million hectares.

Total needed = 122 million hectares.

Our final result then is:

Saved = 170 million hectares (or, possibly, 240 million hectares)
Needed = 122 million hectares

Overall reduction = 48 million hectares (or 118 million hectares if it really is the case that 50% of the area under coarse grains is used to grow feed). This represents a total reduction of all cropland area of some 3% (7%). Is this at all close? Well, a recent study concluded that if all the existing cropland area was turned over to human food, we could feed another 4 billion people. On the basis of my numbers above, 4 billion people need about 29 million hectares of foodcrops for protein plus perhaps about the same again for the balance of their calories. That comes to somewhere around 60 million hectares, so it looks like my numbers are in the ballpark.

All of this said, it’s worth noting I used a recommended daily intake value for protein requirements. Globally, people tend to over-consume protein when they can get it and as the world’s population gains better access to food, overall protein consumption is increasing. Would it be any better under a plant-based agricultural system? Sadly, I suspect not, so the figure for land needed above could undersell the true situation by as much as 50-100%. If it turned out I have under-estimated by say 50%, then land needed would be around 180 million hectares leaving us in balance – no area gained or lost.

What have we learned? That the vast areas under monoculture crops today aren’t the fault of plant-based diets but are rather due to the system we have and the human demand that we have. Replacing the global diet with plants alone could reduce that area by somewhere between 50 and 120 million hectares but that does seem to depend on people not over-consuming as they do right now. The take away from this might be that replacing all the animals we eat with plants may not make as much difference as people think, at least not insofar as the problem of large areas of monoculture cropping. Something else would be needed.