Should I Stop Eating Meat? Part II

Beginning to drill down into some CO2e data to bring out a few conclusions

7 min readSep 23, 2021

Part of a series. Part I here.

How on earth are we going to gain traction on this when we know nothing about agriculture?

Just like everything previous, the aim is to keep everything as simple as possible with complexity only added when necessary. On that note, instead of having to come up with our own estimates of how the agriculture GHG numbers break down, we can continue to rely on yet more cleverer people and apply a dusting of sanity checking as we go to e.g. ensure new work roughly lines up with what we’ve already established — that agri is around 25% of the problem.

This paper by Poore and Nemecek (2018) — from now on PN2018 — will form the basis of our estimates and can be independently checked by this more recent paper by Crippa et al. (2021) and also this paper released only a few days ago. They all provide an estimate of GHG emissions from agriculture with PN2018 breaking this down by food product, but first let’s have a look at the aggregated pictures to see if their independent analyses line up with what we’ve established so far.

Drawing — Image by Hannah Ritchie at Our World In Data

Again, leveraging the work that HR has done at OWID, we can compare the PN2018 aggregated numbers side-by-side with the Crippa et al. numbers. What can we see:

the numbers roughly line up (even when dis-aggregated into sub-sections)
they come out at around 25–30% of GHG emissions which is higher but largely agrees with what we’ve seen so far for agriculture
the differences between the numbers can be explained through slight differences in how they’ve achieved their results

So all in all things are looking pretty good:

we have some data we can use to attempt to gain insight into whether or not certain foods are generating more GHG
that data (when aggregated) also seems to line up with what we have seen before from independent sources for agriculture’s total footprint

One final word on what PN2018 have done as we will be using their numbers a lot to start with and it’s worthwhile understanding how they have been generated. Their study is considered one of the most comprehensive ever done and is a ‘meta-analysis’ that spans almost 40,000 farms/food processors in 119 countries by combining the work of many other researchers. You can have a flick through the data here on a Google doc if it tickles your fancy (because it sure tickles mine).

The method used is called a ‘Life Cycle Analysis’ (LCA) and takes into account everything at all stages of food production. Using the example of a car, instead of just measuring the GHG emissions of driving the car, they measure all emissions related to driving, manufacturing and supplying the car along with mining the materials for it. This ensures we include ‘indirect emissions’ and get a better perspective of the footprint of an activity — rather than just the ‘toe print’ as the OG of carbon footprinting Mike Berners-Lee puts it. More from him later.

But without understanding how this breaks down by food product, I have no idea what I can do

Agreed. But it’s not even just that, it’s actually a load more complicated than that. If we’re going the whole hog then really we need to worry about:

food is not just calories, we need to look at more than just that to prevent micro nutrient deficiencies (‘hidden hunger’)
we’re part of a global world, but supply chains vary substantially depending on where you are in that world — e.g. does what I choose to eat in the UK really impact Brazilian deforestation?
is it even safe to not eat meat?
are there not huge variations in GHG emissions between countries and even within countries due to farming practices?

And all of these questions matter. However to attempt to try to gain some traction like every (overly indoctrinated) economist let’s:

first make a few (unrealistic) assumptions
generate a few conclusions that only hold under these assumptions
start peeling back the assumptions
check if the conclusions we generated still hold

So first, the assumptions

Assumption 1: It is safe to not eat meat

Let’s assume that meat is not required for a diet that provides all the macro and micro nutrients required to live your life — whether that’s a the life of a child, adult or pro athlete.

Assumption 2: Meat can be reduced to calorie and protein intake

Following on from above, we can consider meat’s role as food to be solely in terms of the calories and protein it provides. A slightly weaker form of this assumption is that other nutrients (in all food, not just meat) are proportional to their protein content. That way, if we substitute non-meat protein sources with equal protein they will have equal supply of those other nutrients.

Assumption 3: Post-farm emissions are equally distributed

Let’s assume that once out the farm gate, the emissions from the ‘supply chain’ stage that constitute processing, transport, packaging and retail are evenly distributed between foods, particularly between animal and non-animal based foods. This removes the need to worry about global numbers masking the fact that a certain type of food might be hugely energy intensive in the processing stage even if very GHG-lite in the farm stage. We will examine this later.

Assumption 4: We eat like ‘average global citizens’

The largest and least true assumption of all, but necessary to make in the beginning. This assumes that we can treat ourselves as ‘global citizens’ and that when we eat, we eat food that is randomly selected from around the world. Under this, we can drastically simplify the problem such that we can base our analysis on global averages of GHG emissions for each food product. This absolves us from having to care about how any particular food has been farmed e.g. grass-fed beef vs ‘Brazilian-rainforested pellet-fed beef’ and instead can hide behind global averages for some nice simple analysis. Neato.

Can we finally start looking at some numbers???

Yeah go on then. Time to get into some numbers and visuals — again courtesy of the work of HR at OWID. Although the western world is completely overly obsessed with protein (see here and here), we will use it as the main method of comparison as that is the predominant nutrient we get from animal-based products and so should compare based on it (rather than calories or weight — no one has ever eaten a steak for its energy content. Although people who eat it solely for its protein content need examining as well).

The above chart shows for a bunch of common ‘protein-rich’ foods the CO2e associated with producing 100g worth of protein from them. To be explicit, the top bar shows that to produce 100g of protein from beef — only reared to be beef (not producing dairy) — the average CO2e produced in the PN2018 study was 49.89kg. Let’s make some beef, sorry brief, observations:

yeah it looks like red meat dominates, particularly when farmed only as beef and not to produce dairy throughout their life. A first look into Nicolette Hahn Niman’s catchphrase “It’s not the cow, it’s the how. Also more on that later (there’s a lot to get to later).
a rough ordering seems to be red meat, dairy, pigs and poultry, fish and then veggie stuff
most well-championed unprocessed ‘veggie alternatives’ are dwarfed by red meat in terms of average CO2e emissions

It’s worth including the same graph but for 1,000 kcals just for completeness but as mentioned it’s not really a fair comparison to compare animal and non-animal products on a kcal basis as that is not generally their function in our food system.

So what can I do with these numbers? What does this actually show and how can I contextualise this?

So far all we’ve shown is this: under the strict assumptions we made above, it appears that due to the non animal-based food stuff (peas, nuts, tofu etc) having much smaller bars in our bar chart than the animal-based food stuff (beef, lamb, chicken etc) that a vegetarian diet has a smaller environmental impact than a meat heavy diet. But this is:

not a useful conclusion (owing to the original unrealistic assumptions)
not very tangible as we don’t measure our lives (yet) in terms of CO2e to compare things we get up to every day in terms of environmental impact

Before relaxing the assumptions we have made, let’s attempt to contextualise the above numbers with the help of Mike Berners-Lee (MBL) and his amazing book How Bad Are Bananas? (HBAB) to answer questions like:

what does it mean that I ‘save’ 23.25kg CO2e by eating 1,000 kcals of tofu rather than beef?
based on various meat-eating starting points, how does becoming a vegetarian compare to other things I get up to each year like flying and buying wicked cool new clothes?

And for that, we need to start getting into the numbers.