Warning: Long post ahead. Here, I am responding to a fairly detailed criticism from a blogger who suggests she has given proof that vegans cause more harm than non-vegans when considered on simple numerical terms. I believe there are flaws in her calculations. My post here is aimed at highlighting those arguments I feel could bear more scrutiny. In the end, I conclude that globally it probably has to be less harmful to eat plants rather than to eat animals (on a strict numbers basis, I am not going to tackle questions of environmental or other impacts). On an individual basis, it might not be so clear and there probably are sound grounds to claim that eating some meat might be better overall. In the end, it is your call. But choosing to adopt vegan ethics means that when you make that call, you really do care about the animals we hurt to live.
Skip to Conclusion for a summary of my argument.
I recently watched a very good video from a Youtuber called Farming Truth. This person also has a website and FB page where she spends a great deal of time and effort examining vegan claims and defending modern day animal farming against those claims.
In this video, Farming Truth argues that it’s now proven – vegans do kill more animals. But is she right?
Video resources: https://farmingtruth.weebly.com/dc-video.html
I must say this is one of the most comprehensively researched and calculated estimates of the comparison between animal deaths in agriculture that I have yet found on the web. However, while there is a wealth of great information and detailed analyses in this video, I am less convinced of claim that it constitutes “proof” of anything at all.
First because I think the numbers aren’t as convincing as portrayed, but also – and perhaps more importantly – because I feel that something so complex cannot be reduced to a simplistic numbers game. I felt the same recently when Matthew Evans wrote his book extolling the virtues of a more natural farming system and I took up the challenge of thinking about the numbers side of his argument (see my thoughts here).
To be honest, I find myself disappointed by such arguments. Are writers who try so hard to criticise veganism missing the point? I think so. In the end, veganism is just the idea that whenever we can we should behave ethically and in accord with our broader moral beliefs in our relationship with other animals. That seems admirable, not something to be derided.
Nonetheless, this question seems to be a perennial favourite, wheeled out from both sides ever since Steven Davis made his argument about crop-related deaths versus grass-fed beef back in 2003. And it IS a good question. Let’s face it, if it really is the case that we do worse to grow crops, shouldn’t we at least be open to thinking about that?
In this blog post, I aim to tackle Farming Truth’s conclusion that vegans always cause more animal harms and deaths than non-vegans.
First, I will summarise Farming Truths argument. Then, I will note some possible errors on her part in terms of the numbers she introduces. Lastly, I will take a shot at recalculating the numbers as fairly as I can and see where that leaves us. At worst, I hope my analysis offers grounds for Farming Truth to re-examine her case.
Farming Truth’s conclusions
Growing crops causes a lot of harms, but in particular, ocean dead zones from fertiliser runoff and Australian mouse plagues cause truly staggering numbers of deaths. Nearly all crops are grown for people, so we can’t discount this harm by claiming that some significant proportion is grown just for animals.
Taking all of this into account we find that averaged globally, approximately 114 animals die per hectare growing crops while just 46 animals die per hectare under pasture.
Considered on several bases – per gram of protein, per calorie and per serve – it turns out that more animals are killed per unit for plant foods than for animal foods. Most notably, Farming Truth concludes that vegans cause – proportionally – 1.16 times more animals to be killed per day for their diet than do omnivores. This latter statistic is particularly telling.
Farming Truth criticises several vegan arguments about the morality of veganism, vegan responses to various claims about crop-related deaths, and holds Mike Archer up as a sort of exemplar of omnivorous reasoning.
Lastly, Farming Truth exposes the matter of wild animal suffering versus farmed animal suffering.
In this section, I want to raise some areas of concern for me in Farming Truth’s analysis. I won’t say my criticisms are right, rather I offer these in the spirit of encouraging further discussion.
Some important numbers are estimates/guesses
Now, it’s important to note before we start that many figures used by Farming Truth are guesses. We do not really know how many wild animals are found in typical areas of crops. The number may be quite high in some sorts of crops in some places (eg mice in wheat fields in Australia) but may be very low in others (eg tomatoes in the field or indoor systems – consider vegetable production in Holland, or fruit and vegetable production in Almeria, Spain). Farming Truth has made some broad assumptions and drawn averages. I suggest we take all these numbers as very uncertain and therefore useful for indicative purposes only.
Interestingly, even though we don’t really have very good empirical data, Farming Truth places a great deal of emphasis on mice killed in mouse plagues as well as sea animals killed from fertiliser runoff from crop farms.
Yes, a lot of mice are killed during mouse plagues. However, mouse plagues these days only seem to occur with any intensity in China and Australia and even then not in all grain growing regions simultaneously so it seems unreasonable to use these numbers in claiming some average across the globe. But I am no statistician, so who knows, perhaps it is fair enough. That said though, plagues in Australia at least are less frequent lately and so the numbers being killed seem to be rather less than is claimed here.
For example, the last really serious plague was in 2010/11. Since then, mouse numbers have been variable. Good growing conditions in 2017 led to an outbreak of high numbers in some regions, but so far in 2019 numbers have been very low due to very dry conditions.
You can see mouse conditions for any period since 2014 here: https://www.feralscan.org.au/mousealert/map.aspx
Fish deaths from hypoxic events
In regard to fish deaths from hypoxic events, I don’t think Farming Truth’s strategy is right. Essentially, she determines the number of fish at risk from such events (using an estimate of global fish populations) and then divides this by the number of global hectares of crop lands to derive the number of fish at risk from crop related activities and then discounts this by 50% to finally arrive at a number of deaths per hectare. Again, this looks to me to be at best a slightly informed guess
There are, after all, a variety of sources responsible for harmful runoffs, including everyday urban and industrial activities, shipping, crops grown for purposes other than food (eg corn for biofuels), flood events and even the meat industry itself (for example, this article in The Guardian).
Here are two sources that offer a range of possible causes for hypoxic events:
Eutrophication in coastal environments
Fish indicators of environmental change
With this in mind, what is a reasonable conclusion? To be honest, anything I come up with would be just as much a guess as Farming Truth’s own guess. The numbers of sea animals dying from such events worldwide may be greater or smaller than Farming Truth claims, but it does seem that we cannot sheet these all home to crop farming for food. It’s a good point, really, but gee, these numbers appear to be little more than a stab in the dark.
Are crops really mostly grown for human consumption?
Farming Truth goes on to claim that nearly all crops grown are for human consumption. Now, this is a constant refrain – vegans say some large proportion of crops are grown to feed animals, farmers say the opposite. I have spent a lot of time researching this and it’s all a bit murky. Certainly crops like wheat and so on ARE grown with human consumption in mind but a considerable proportion never makes it to that purpose, perhaps due to not meeting the grade. A LOT of wheat and similar grains really service animal feed markets (currently in Australia, some 60% of domestic demand for wheat comes from the animal feed industry). As well, some industry sources suggest that there ARE grains and cereals grown purely for feed.
One crop that gets a lot of attention is soy. Vegans tend to claim that most soy is grown for animal feed, but this isn’t strictly true. While numbers vary, generally speaking direct human consumption accounts for around 10-15% of global soy yields. The rest – 85-90% of global yields – is used to produce oil and animal feed (meal).
You may have seen figures suggesting that 80% of soy grown is used to feed animals and the other 20% is used for oil but this shows a misunderstanding of the process. In fact, pretty much ALL of the soy NOT grown for direct human consumption is crushed for oil and the oil is almost all used in human applications such as cooking oil, margarine, food packing and biofuels. The residue from the crushing process is then used to produce meal for animal feed. It’s an attractive feed stock due to its high protein content. In other words, 80-90% of all soy grown produces oil for human use and meal for animal feed. This is why farmers say that most soy is grown for human use.
However, the demand for animal feed substantially drives soy production with more farmers globally adopting the crop due to its high returns. Soy meal has a huge global market. Again, using Australia as an example, we see soy production falling off since about 1990 in tonnage terms, largely because the focus has moved to servicing human food markets. However, imports of soy meal as animal feed have surged since then and in 2018/2019 we imported around one million tonnes (by way of comparison, Australia grows just 20-40,000 tonnes of soybeans annually). By the way, I should point out that soybean meal production in China totals around 67 million tonnes annually while in the US it is around 45 million tonnes. In Brazil and Argentina it is around 33 million tonnes annually. That’s a lot. My guess is that this tracks livestock production figures, in particular pork and poultry.
Around the world there is active research on finding more uses for soybean oil in order to maximise the return for the increasingly large volumes of soy grown. In fact, soybean oil is one of the most common cooking oils and is increasingly used in the food industry for that purpose. Interestingly, it may be that the meat industry is driving both demand for soyben meal AND oil, given that a lot of the cooking/frying in the food industry is cooking meat!
I believe that it is therefore somewhat misleading to claim that crops used to provide animal feed are really grown just because people eat them. It is almost certainly the case that were there no animal farming, there would be far less grains and the like grown.
How much protein is there in plants?
In the accompanying spreadsheet, Farming Truth uses a protein value in grams per pound to make some important calculations. However, I think these numbers are incorrect. Here are the numbers used, followed by my updated numbers derived from the SELF Nutrition data website.
Food grams of protein per pound yield
Number of wild animals killed growing crops
Farming Truth aims to determine how many animals are killed on average to raise a hectare of crops. The basis for the estimate is the fact that on average we should find 111 animals per hectare living on croplands. This is simply a guess on her part and has absolutely no empirical basis whatsoever. We should bear that in mind.
To ascertain the death toll on these croplands, Farming Truth then assumes on no obvious basis that all activities other than harvesting cause 30% of those animals to die (calling this a “lowball” estimate!!). Again, this is a guess. Add to this the 60% from harvesting as assumed by Davis and Farming Truth concludes that growing crops causes the deaths of as many as 90% of the local wild population, or 100 animals per hectare.
As supporting evidence she refers to an anecdote from an Oregon farmer who noted all the small animals were killed when he harvested his crops, even though we have no mention of the actual number this represents in this real world example.
Should we accept these numbers? I don’t think so. I draw attention to a blog started by a wheat and sheep farmer in the UK who maintains that many animals are killed in cropping, especially wheat. I have spoken to this fellow quite a bit and he remains adamant that he is correct. Yet, he aimed to collect evidence for the large number of wild animals killed and to document this on his blog over the course of a full season. The end result? Not one animal death of the kind we are talking about was recorded.
That isn’t to say there were none, but here is someone highly motivated to find considerable harms caused to wild animals from his activities and yet he was unable to do so. Yes, he did show that insects, worms and snails died, but generally speaking these animals aren’t accorded the same level of moral responsibility as animals like rabbits and birds. So right there is an anecdote that illustrates a quite different outcome from that of our Oregon farmer. http://theydiedforyourbread.blogspot.com/2018/
Can we offer any sound evidence for the numbers of animals that ARE killed in cropping activities? I don’t think so. Last year, Bob Fischer and Andy Lamey published a paper, “Field Deaths in Plant Agriculture“, in which they sought to scour the literature and determine whether it is possible to estimate the toll in cropping. I won’t spend much time on this research save to note that it tends to confirm what I am saying – we just do not know. On the whole, it does seem numbers aren’t as high as many suspect.
Fischer and Lamey are unable to offer any clear estimate, noting the paucity of solid empirical data and the fact that estimates for one species in one context don’t necessarily extrapolate across all species and contexts.
Nonetheless, they do make one relevant point about mouse plagues in Australia. Mice are present during plagues across different agricultural contexts and damage agricultural equipment and installations in both plant and animal agriculture. Farmers of all persuasions kill mice during plagues, so it seems a little unreasonable to place the toll of mice killed squarely on crop farming activities (mind you mouse numbers in plagues seem not really relevant given that Farming Truth guessed how many animals are killed on croplands using as a basis a proportion of estimated global populations of all animals generally).
On the whole, I come back to this central point. In the absence of good empirical data, any estimates amount to guesses. We just don’t know how many animals really are killed to grow crops, at least not with sufficient detail to make any compelling global estimates.
I do have one caveat however: vegans tend to regard insects as sentient beings and hence deserving of similar prohibitions against harm as other animals such as mammals. If we really are to worry about insect deaths in cropping, then there is no contest – raising free range cattle is probably about the least harmful thing we can do for food.
Comparing like for like: numbers of animals killed for food
In Part 4, Farming Truth takes the number she has derived for crop deaths (114/hectare annually) and compares it to the number of farmed animals killed per hectare of land grazed (46/hectare annually) to conclude that crop farming causes something like 2.5 times as many animals to die as does animal farming. I don’t know about this. To me, it seems that in working out animals killed per hectare in cropping, we are really seeking some metric for calculating the total number of animals killed to grow crops. Obviously, when it comes to the numbers of farmed animals killed, we can know this directly without resorting to some calculation. This number just is what it is and hectares don’t come into it. In other words, in both cases the aim is the same – what is the number of animals killed? For farmed animals, we can count them directly. For crop related deaths, we have to infer the number from some rate (which, incidentally, remains a guess).
Consider too that Farming Truth estimates a global value for animals killed in cropping but then uses a US number for animals directly killed for food. Curiously, she also ignores sea animals. So, shouldn’t we simply compare apples with apples and estimate the global totals? By Farming Truth’s estimate, some 1.5 trillion animals are killed for crops globally. On the other hand, what estimates I can find suggest that each year we kill as many 70-80 billion land animals and a further 1-3 trillion fish. This latter number doesn’t include other sea animals such as squid, octopus, lobsters and crabs and bycatch. I shouldn’t be surprised if the total number is not more than 3 trillion each year. It seems quite possible that twice as many animals are killed for food as are killed in growing crops, at least on Farming Truth’s estimates for crop related deaths.
This same concern arises in Part 5 where Farming Truth calculates the numbers of animals dying for each kind of agriculture on several bases – per 100,000 calories, per 10,000 grams of protein, and per 1,000 serves. I am no mathematician, but I’m not sure this is a valid way to approach this problem. Again, this seems only to be striking a rate. Rates are important, sure, but if the aim is least harm, we really should be more interested in total numbers. Rates seem to tell us how quickly we reach a total, while the total is what counts when taken overall. Nonetheless, I’d like to see the calculation corrected by adding in sea animals. This helps us to determine where the worst harms occur.
In similar fashion, is assessing deaths as some average per serving the right way to consider this? As I note above, when we eat animals for food there is no average needed – we just eat some number of animals. Consider that the typical omnivore must eat a particular number of animals each year. The typical Westerner eats beef, chicken, pork, eggs and fish and most estimates I have seen suggest that this results in as many as 80 to 100 animals per year (I estimate I caused as many as 150 animals to die each year for my diet when I ate meat and one website I visited claimed we kill as many as 300 sea animals per person each year for food). Yet, on the numbers Farming Truth is using it seems that the number is around 15. Something is wrong here.
Something that stands out for me, having been digging into this stuff for a little while, is that the yields used in calculating plant related deaths seem off. In the related spreadsheet, Farming Truth has used these yield values:
These seem very low to me. I found this site, Factfish, that gives details of global values for a range of indicators. Using the averaged global crop yield values from Factfish gives me these numbers:
I’d like to see justification for the lower values used by Farming Truth. If the higher values are correct, what effect does that have on calculations?
A different way to think about this
In the end, exercising vegan ethics is an individual choice. While the broad concept of veganism goes beyond food, it is in our everyday diets that most people might intersect with the philosophy. In terms of evaluating the impact of our food choices, I think the calculation is ours alone. In the following, I want to suggest a way for an individual to consider whether the animals killed raising crops might be an influencer on their decision. I am not going to try to extend that to some generalised global case.
Globally there are way too many local matters to be considered – I can’t hope to make any meaningful generalisations. Beyond of course that I believe we want to do best. Whatever that is.
So, here I try to strike a solution to the problem of animals killed in raising crops versus animals killed for food such that someone in Australia, for example, could respond ethically.
In her calculations about proportional impacts, Farming Truth has used an idealised number of servings per day to establish a kind of common basis for making the least harm calculation. I think this is a sound strategy. Of course, people eat a remarkably diverse range of foods and there are all sorts of diets, so to an extent this is just a sort of abstraction but the idea seems sound.
So, let’s follow Farming Truth’s concept of setting a generalised standard. Farming Truth has used as a guide 15 serves of plant foods and 5 serves of meat and dairy per day (drawn from the ‘Food Pyramid’ dietary guidelines).
Now, this leads us to an important point. That is, all of us, vegans and non-vegans would, on this idealised model, eat plants. Non-vegans eat 15 serves of plants each day. So do vegans. This means that all of the crop related deaths to grow THAT food falls on everyone’s shoulders, not just vegans. To an extent, we can disregard those deaths, though in a vegan model we’d seek to reduce those as well. To be fair, Farming Truth has taken this into account in one of her calculations where she finds that proportionally, the vegan diet causes 1.16 times as many animals to die for a vegan’s daily food.
With this in mind, I want to propose that in fact, the number we want to quantify is the number of animals killed to grow those crops needed to replace all of the animals in our idealised diet (ie the five serves of meat and dairy per day that non-vegans should eat). That is, if in the idealised model we are all eating 15 serves of plants then the deaths from that production is a shared cost and irrelevant in deducing whether vegans or omnis cause more harm. What matters when evaluating the harm from a specifically vegan diet is whether killing animals for those five serves of meat and dairy leads to a greater or lesser death toll than replacing them with plants.
According to Farming Truth’s calculations, on average five serves of meat/dairy delivers about 90 grams of protein daily. Annually, that is about 33 kilograms of protein. In Australia, I believe the average daily protein intake is about 110g. Presumably some of that comes from non-animal sources, so I think 90g is a reasonable benchmark.
Replacing that protein with plants would mean eating high protein crops (though perhaps not so much wheat – I think we already may be eating about as much wheat as we need). Let’s assume as representative the following: chickpeas, black beans, lentils and soybeans (eg tofu). While vegans might drink various plant milks rather than dairy milk, the protein content varies from negligible to some, so I think we’d have to discount plant milks as serious protein replacements. While various foods can be fortified with protein (and B12, calcium, iodine etc), I think we should leave that out for now as an uncontrolled variable.
According to SELF Nutrition Data, these foods offer the following protein per kilogram:
Black Beans 216
Soy (as tofu) 160
Let’s average that out to 205 grams/kg. This means protein content is around 20% of source product. That suggests we’d need around 160 kg of those crops annually to deliver the equivalent amount of protein as five serves of meat and dairy (ie the 33 kg annually). At an average yield of 2100 kg hectare, that translates into about .08 of a hectare. At 114 wild animals killed per hectare, the vegan diet causes an extra nine animals per year to die over and above the shared baseline for someone eschewing meat and dairy.
Referring back to my earlier estimate of the number of animals killed per year for food for non-vegans , about 80-100, it seems the vegan diet causes fewer animals to die.
Yes, it is quite possible, IF those numbers are at all right, for the average non-vegan to do better by choosing to eat fewer animals, for example by not eating any sea food and perhaps limiting chicken meat. In this kind of scenario, our ethical omnivore might eat only range grazed beef and lamb.
However, that choice being less harmful turns on the number of wild animal deaths from growing crops being 114 per hectare which, as we have seen, is something of a guess. It might be more or less, though the evidence suggests it may be considerably less. If it turns out that fewer animals are killed to grow these kinds of crops, it might be quite different. Consider if the number is perhaps 40/hectare/year. Then, the vegan’s toll might be as few as three. I think it would take a fair effort from the average non-vegan to keep their animal toll down to that few. Not impossible, though.
The bottom line seems to be that most of us probably will do least harm by eating a plant-based (ie vegan) diet. Still, someone who takes all of this into account but still wishes to eat meat could choose to eat carefully so that they minimise the number of animals killed for their food. Eating just a few range grazed animals probably is ethically defensible, at least on a numbers basis.
On a purely numbers basis then, this suggests that the best things we can do to reduce the harm from our eating choices are to increase the quantity of plants in our diet (preferably eat a plants-only diet), substantially reduce or eliminate our consumption of fish and chicken and eat mainly or only range grazed beef and lamb.
I’m not convinced of the value – in the absence of genuine empirical data – of lengthy calculations and anecdotal claims in an effort to discredit veganism. It might be true that more small animals die to grow crops than to raise range grazed beef. But it remains a fact that we don’t just eat beef – we farm, commodify, harm and kill very many animals for food, fun and profit.
Veganism, at least as far as I am concerned, is the idea that we actually give a damn about this. We don’t just ignore any moral duty to the animals we choose to treat that way. Similarly and perhaps just as importantly, we DO owe a duty to the millions of animals we kill incidentally – the sea food by-catch, the pest animals, the vast numbers of insects and the innocent bystanders. That duty demands that we care.
Insofar as Farming Truth’s calculations go, I think her case is flawed. The issue is important; we should care that a great many animals might be killed or suffer terribly for us to grow crops to eat. But the evidence is sketchy.
I’d like to see Farming Truth’s analysis revisited and my several concerns above addressed – the lack of empirical evidence for the 114 deaths per hectare per year in protein crops, the fact that sea animals have not been included, the discrepancy in crop yield values and whether relative protein content is as claimed.
What outcome might be reached? On the evidence I think it’s very hard to dispute the fact that globally it probably is the case that IF we could do so, it’d be very much better not to kill animls for food. Taken individually in countries like Australia, it probably is the case too that for the average person, it is likely to be better to eat plants rather than animals. But it does seem that there is the strong possibility that one could choose to include some animals in one’s diet and be confident of doing less harm than otherwise.
In the end, knowing more about how our food is produced and caring about the systems used and their impact on farmed and wild animals as well as local ecologies is important. These facts should count for a lot more when making our food choices.
Finally, I want to ask that we don’t just dismiss veganism as a fad diet or a straight-jacketing binary choice. I think the issue is at once more complex and yet demanding of genuine consideration by all of us. Because in the end, it just asks us that we care.
Shouldn’t we all want that?