Note: this post was put together on a whim on a Friday afternoon; the analysis is woefully incomplete. I hope to update it, though my historical track record on updating blog posts and personal projects is remarkably terrible. That said, if you have any relevant information or insights, please reach out!

Introduction

Roughly daily, I find myself buying food, whether in a grocery store or a restaurant. Therefore, roughly daily I experience crippling uncertainty about how to balance my various objectives which include pleasure, health, affordability, and relevantly for this post, environmental impact. Given that I am a car-free urban apartment dweller, most of my carbon footprint¹ comes from airplanes and food. It’s pretty obvious what to do about the former: fly less.²

The food is more complicated. There are lots of ways food production can be bad for the environment - carbon emissions, nitrogen run-off, deforestation, land use, etc. And of course there are significant moral concerns about the welfare of agricultural workers and farm animals. Because food choices are complicated, we all use heuristics, though a common heuristic seems to be “I eat what is tasty and cheap and easy to acquire.” I perceive that many (most?) people who share my concerns about food externalities use the heuristic “be vegetarian.” That’s an okay but imperfect heuristic. For example, eggs are probably worse for animal welfare than beef³, dairy has more environmental impact than potatoes, and responsibly farmed bivalves are probably a quite low-impact source of protein (and delicious!). I wonder whether fried foods have an outsized environmental impact and thus whether a more complicated but optimal heuristic than “be vegetarian” would penalize fried foods. I haven’t read about this before and a super-quick Google search didn’t turn up anything, so here we are.

Back of the envelope arithmetic

The problem with fried foods is that most of oil used for frying ends up being discarded, not eaten. But that oil contains a lot of calories, and growing those calories has an environmental cost in land and fertilizer used, fuel burned, etc. Is this effect significant? I don’t really know, but considering some home-fried french fries is an easy place to start:

• There’s a 1-quart bottle of vegetable oil in my kitchen. In round numbers, it contains 8000 calories.
• That oil is enough to fry in a small pot which can probably fit two potatoes worth of french fries. I can probably use the oil 4 times before it needs to be changed, for a total of 8 potatoes of fries. Let’s round up (reducing the estimated environmental impact of frying) and say that oil is good for 10 potatoes.
• calorieking.com says that 10 russet potatoes is ~1600 calories. Let’s assume that when a potato is fried, the weight of water lost equals the weight of oil absorbed. I’d guess less oil is absorbed - french fries feel less dense than raw potatoes - but this is also rounding in fried foods’ favor.
• An equivalent weight of french fries is something like 4800 calories, for an increase of 3200 calories from oil absorbed.
• That means that when the fry oil is discarded post frying, at least 60% of the calories in the oil are being discarded. Or, the combined “cooking yield” of 8000 calories of oil and 1600 calories of potatoes is only 50%.
• So as a very rough approximation, what happens if we double the impact of potatoes? This StackExchange answer (gives some rough sources) puts potatoes at about 1.5g CO2 per calorie. If we hand-wave away the differences between growing potatoes and seeds for oil, differences in post-farm production and storage costs, etc., we can estimate that french fries are around 3g CO2 per calorie, which is comparable to eggs, milk, and chicken. Interestingly, this article includes both production and cooking, etc. related impacts and gets pretty close to that 3g CO2/cal number for potatoes. Maybe that’s for fried potatoes? Or maybe that’s just storage etc. for potatoes; I think they’re held in artificially humidified warehouses, which sounds carbon intensive.
• This website which I have never heard of (but hey, easily digestible data…) puts potatoes at 1300 calories per square meter of land (truly kcal, but I continue to use little-c-calories for convenience) and rapeseed (canola) at about 750 calories per square meter. So those 4800 calories of french fries took about 12 square meters of land while a pile of potatoes with equivalent caloric content would take less than 4 square meters. So in land use the efficiency penalty is perhaps ~3x? That’s probably too unfair to fried foods as potatoes are a quite land-efficient crop, and the same site puts other oils at twice the caloric density as rapeseed.

Factors not considered

• What if I fry something besides a potato? Probably the math is even worse. French fries are porous and have high surface area to volume, so they can absorb lots of oil, meaning fewer calories into the trash when the oil is discarded. Potatoes are a very efficient staple crop, so if some other food is fried, the relative efficiency may not be as bad. But it’s not clear this should be a relief if the other food to be fried starts with higher impacts than potatoes. An abbreviated analysis on doughnuts: a Krispy Kreme glazed doughnut is ~200cal, half from fat. Assume all that fat is from fry oil and that 1/4 of total fry oil is absorbed into doughnuts (probably a lot less is). Then 1000 calories of oil and 250 calories of dough give 500 calories of doughnut for a total “caloric yield” which is order-of-magnitude similar to the fried potatoes.
• How does frying at home differ from a restaurant setting? Restaurant fryers are much bigger than my little pot, and presumably better at avoiding oil breakdown. (E.g. frying in a pot on a stove is bad because the bits of food that fall into the oil sink to the bottom and are then exposed to the highest heat in the cooking vessel, and at high heats these bits of detritus accelerate oil decomposition.) On the other hand, my home pot has a high fried food to volume ratio, probably allowing for more oil to be absorbed. Update 2017-11-26: My brother points out that many restaurants don’t simply throw away their used fry oil. If the used oil is converted to a biofuel (let’s even suppose 100% efficiently) and that biofuel is used in place of some other fossil fuel, some emissions are averted. The magnitude of the effect depends on the efficiency of the agricultural inputs -> plants -> fry oil -> biofuel process. If this process is inefficient, than the energy content of the final biofuel represents only a slim fraction of the inputs to the system, and using vs throwing away this final energy content doesn’t much affect the system’s waste. If the process is super efficient, then fry oil as biofuel could really help. Unfortunately, our prior should be closer to “the process is inefficient.” It’s easy to find order-of-magnitude different estimates for the ratio of energy in to calories out of American farming, but most estimates are well north of 1. That is, producing say 1000 calories of fry oil takes more than 1000 calories of fossil fuels. So very roughly speaking, maybe in a restaurant that recycles its fry oil, fried foods are 1.5x “as bad” as unfried equivalents rather than 2x as bad.
• Bottles of oil don’t grow on plants. Is producing oil from seeds/nuts/etc. energy intensive? Does it also lead to caloric loss? If so, the arithmetic gets worse.
• Individual farming practices matter a lot. Averages are misleading. Food is complicated.

Tentative conclusions

There are a huge number of variables and unknowns. Life-cycle analyses are hard, especially for me given that this is not my job! Nonetheless, a few back of the envelope calculations suggest that frying results in at least 50% of the calories in oil + fried food being discarded. Since I haven’t read that oils are orders of magnitude more efficient to produce than other foods, this makes me think that as a rough heuristic, we should consider a fried vegan item twice as environmentally intensive as the unfried version. This puts commonly fried foods in the ballpark of yogurt, eggs, chicken, fish - not beef-bad or lamb-bad, for sure, but also not oats-good.

I don’t mean to say we shouldn’t eat fried food; even if my analysis is ballpark correct, it’s still plausibly a reasonable conversion of calories to pleasure. If I were king, we’d have a $50/ton carbon tax, eat doughnuts for breakfast, and worry less.