Lies, damn lies, and [tweeted] statistics

Like a lot of writers, George Monbiot really likes out of context stats. Numbers seem authoritative and incontrovertible. Though more often than not, writers who cite such numbers have little understanding of how such statistics are or were derived. Not sure where George got his latest figure regarding sheep consumption only accounting for 1.2% of the diet (I assume in the UK), as noted in this recent BBC article. He seems to cherry pick a lot of his numbers from Oxford studies and reports.

Regardless, here just below is a breakdown of the primary sources of proteins in the UK diet as derived from FAOSTAT (this is the source for another factoid George cites from a FCRN report noted further below).  First note that “grams of protein” is a meaningless stat. Why? Because all proteins are comprised of different amino acids including the nine essential amino acids. Like a rose is not a rose, a protein is not a protein. So proteins from different sources of food are not equivalent.

The numbers shown on the lines in yellow are the primary sources of protein in grams per day. As you’ll note in the “average” UK diet on a daily basis, people get most of their protein from wheat, milk, and chicken. And yes, doing the math, George is approximately correct at least for protein, 1.5 “grams of protein” from lamb comes out to be around 1.5 to 2 percent of total “grams of protein.”

The lines in green and white are the essential amino acids of each protein source. The first pink column on the left is the RDA (recommended daily amount) for each of these essential amino acids. The middle pink column is the total amount for each essential amino acid from plant sources of protein, and the pink column on the left is the total amount of each essential amino acid from animal sources of protein. As you’ll see, animal proteins are much more nutrient dense. Plus that 1.51 daily grams of lamb has more than two to three times the amount of each essential amino acid as all the vegetable sources of these amino acids excluding wheat.  (Note too industrially grown mono-cropped wheat is grown with a lot of synthetic nitrogen to up its protein content).

FAOSTAT nutritional data is available for every year from 1961 to 2013. It would be great to get data from long before 1960, since that would be even more revealing about how dietary patterns have changed with industrial seed oils, soy bean oil, processed foods, and the huge increase of chicken consumption. Per the 1961 data, people in the UK ate only 2.51 grams of chicken per day, so chicken consumption has gone up over 600 percent. Lamb consumption in 1961 was 3.91 grams or over 250 percent more than it is currently.  Beef consumption has come down, and pork consumption is about the same. Wheat consumption is also similar.

Now the obvious big problem with all of that chicken, especially in the EU and UK is that those chickens aren’t eating insects on pasture. No they’re eating soy meal in CAFO’s.  About twenty to thirty years after the soy bean crushing industry came to be in the 1920’s, a solution was needed as to what to do with all the left over meal or cake from crushing the soy beans for oil. That solution that came in the late 1940’s was to feed all that soy meal to livestock in confined animal feeding operations particularly to monogastric livestock, chickens and pigs.  Today about ninety percent of all soy is crushed for soy bean oil, and around seventy five percent of the resulting soy meal is fed to chickens. Much of the soy grown in Brazil goes to support Brazil’s huge chicken industry. Brazil by weight produces around 20% more chicken than beef. Brazil’s beef is primarily grass fed and finished (feedlot capacity is only around 2.5 million to 5 million head). A lot of the soy meal from Brazil is also exported to China, the EU and the UK to feed CAFO chickens and pigs in these regions. In the US and globally, chickens also account for the vast majority of land animals killed. One figure I recall from 2014 was that in the US of the 9.2 billion animals slaughtered annually, 8.8 billion were chickens.  So yes, we should eat less chicken.

The fats we consume and use for cooking have changed as well.  Before 1920, almost all the fats humans consume in many regions of the world were animals fats. By the 1950’s only half the amount of fats consumed were animal fats, and the balance had been replaced by industrial and seed oils. By the 1980’s, seed and soy bean oil consumption had risen to around 75 to 80 percent of fat consumption as show in this graph borrowed from author Nina Teicholz.

Here just below is a graph for changes in fats consumption for the United States derived from the FAOSTAT data for the period from 1961 to 2013. In 2013, people on average got 59.58 grams of fat from soybean oil per day,  5.18 grams of fat from canola oil per day, and  5.89 grams of fat from corn oil per day. That translates to 636 calories per day of 25% of a 2500 calorie per day diet. People are obviously not guzzling down these fats from the bottle, but these fats are in fryers and in processed foods that are part of the Standard American Diet [SAD]. So next time, some vegan claims that commodity crops are just used to feed livestock, you can remind that person of these figures, plus tell them where bio-fuels and ethanol come from. Commodity crops, like corn and soy, have multiple co-products and by-products including seed oils, soy bean oil, ethanol, alcohol, animal feeds, industrial solvents, biodegradable plastics, biofuel, processed foods, etc.

Here just below is another graph for changes in fats consumption for the United Kingdom derived from the FAOSTAT data for the period from 1961 to 2013. In 2013, people on average got 11.94 grams of fat from soybean oil per day,  24.07 grams of fat from canola oil per day, and  2.71 grams of fat from palm oil per day. That translates to 350 calories per day of 14% of a 2500 calorie per day diet. So not quite as bad as the United States that devotes nearly 180 million acres to growing either corn or soy as part of the United States bio-based economy supported in large part by crop insurance.

The transition from animal fats to vegetable oils hasn’t been the only change that has occurred in modern diets of developed countries like the US and UK. Sugar, and another commodity crop derived co-product, high fructose corn syrup, has gone way up as well. So too have heart disease, type 2 diabetes and obesity. And though in developed countries, animal fat consumption and red meat consumption have gone down, many especially in the UK want to place the blame on red meat as the primary cause of health issues.

Source: Nina Teicholz

Many of these people in the UK seem to be affiliated with Oxford University , and get funding from the likes of Nestle while collaborating with people like Walter Willett who, while at Harvard, received a lot of research funding from Unilever. Guess what’s in the products of Unilever brands? You betcha, industrial seed and soy bean oils. One recent study from Oxford University (Springman et al 2018), that got a lot of headlines in the news as a rational for taxing meat, relied in large part upon the very weak relative risk [R/R] association WHO cited in their 2015 meta-analysis. The conclusions of that meta-analysis relied on just a handful of epidemiological studies that couldn’t really correct for numerous confounders.

The conclusion? A relative risk association of 1.17 or 17% for red meat and a R/R for processed meat of 1.18 or 18%. The associated absolute risk [A/R] for both red and processed meat was less than one percent. Need a refresher on the difference between absolute and relative risk, click here:  Absolute versus relative risk- What’s the difference?  Remember too that the associated relative risk between smoking and lung cancer is 2500 to 3500 percent. Once you understand what relative risk numbers mean and how they’re derived, you also begin to understand that researchers such as Springman and others at Oxford are really advocates masquerading as scientists with their sensationalized fear mongering to push their dietary religion. 

Okay, so maybe switching to vegetable fats, HFCS and meats reliant upon commodity crops (CAFO pigs and chickens) over the past fifty to sixty years wasn’t such a great thing especially in regards to one’s health and the health of the environment.  This obvious observation, brings me to another factoid Mr. Monbiot decided to share with me via twitter in an exchange this past January in response to my article, George Monbiot once again tilting at the wrong windmill.

This factoid I was able to later track down. Its source is the  Oxford University FCRN’s report Grazed and Confused, that strangely didn’t realize carbon utilization, respiration and sequestration is largely driven by soil microbiology (see- It’s the Soil Biology Stupid). In this FCRN report, the number George cites is just given without any explanation other than referencing the FAOSTAT’s. How the raw data from FAOSTAT was parsed I don’t know. I strongly suspect George doesn’t know either.  So needless to say, I  did some of my own research to try to ascertain how the number was derived as well as whether or not the number is even relevant.

Let’s take George’s tweeted number at face value. First regarding land use, land isn’t interchangeable. So, as noted in Mottet et al 2017, a lot of land used for grazing can’t be used for crop use. This is especially true of Africa, Australia and the western and northern half of North America. Topography, season length, amount of rain fall, soil type, etc all impact whether land is arable or not. Most grasses, especially perennial grasses, require a lot less rain (green water) than vegetables, fruits or grains. Many crops require irrigation (blue water). Living in more humid environments, many people in Europe and the Eastern half of the United States have distorted perspectives. The entire planet isn’t like the UK or the Eastern half of the US.

Such a globalist, or dare I say colonialist, perspective is also a failing of using the global number George chose to present. There’s a lot of variability on sources of protein consumption on a regional basis. In Australia for example between beef and dairy per the FAO stats, Australians get over 30 percent of their “grams of protein” from what are largely pasture based cows and beef cattle.  Stats also get distorted by huge numbers of certain populations who get most of their “grams of protein” from low nutrient density rice or wheat as well as from grain fed monogastric livestock. For example, 1.2 billion Chinese and many in India get  very little protein from pastured meat but a lot from rice or wheat.  Then, George’s single gram number also is severely distorted by methods of production, since most livestock “globally” are in integrated systems or ones where a considerable portion of a ruminant’s life is spent on pasture or range land eating forage (grasses, forbs, legumes) before being “finished” in a feedlot. That’s certainly the case here in the United States, where most (per USDA statistics 68 million out of 84 million head) of the beef cattle inventory (cows, bulls, replacement heifers, calves, stockers) lives on grass in cow calf and stocker operations, before yearlings get transferred to feedlots where these yearlings are fed a lot of roughage (like silage, a grass) before being transitioned to primarily concentrates (grains, and spent grains) during the last month or two of their lives before slaughter.

Plant and land animal proteins per country – numbers from FAOSTAT’s – excludes seafood

I didn’t compile the 1961 numbers, but those too would be very illustrative since they’d be more red meat heavy. Again the transition to grain fed and finished sources of animal protein as well as to vegetables fats, as previously noted, has been a relatively recent phenomena begot in large part by the seed and soy bean crushing oil industry . This transition has been at a great detriment to our health. Wonder why, Oxford isn’t doing studies suggesting taxes on industrial vegetable oils. Oh wait, never mind, that may upset their benefactors and overlords like Nestle and Unilever.

Needless to say,  in many  countries and regions, much larger percentages of protein are coming from pastured, range land sources in part or in whole than George’s global stat would seem to reveal. We’d also be better off with more of our protein being produced in regenerative systems including integrated systems like those of Gabe Brown, Colin Seis, Bruce Maynard and others where meat and vegetables are produced on the SAME land in stacked enterprises. Part of any argument, especially a vegan argument, is transitioning to systems of food consumption and production we or they would like to exist.  Therefore what’s we or they would like to see isn’t the current prevailing norm. For George and his Oxford cohort, their vision seems to include a lot more phytoestrogen full goitrogenic mono cropped soy beans grown with synthetic fertilizers and pesticides to be consumed in hyper processed foods or fed as amino acids to growing stem cells in energy intensive bioreactors. Not exactly a vision, I share.

That brings me to one last stat,  again emanating  from Oxford via the Guardian, but this time not via George.  This was yet another global  “grams of protein” argument about land use and goes as follows:  Meat produces only 18% of calories and 37% of protein, yet uses the vast majority – 83% of farmland.  These numbers come from the 2018 meta-analysis by vegan lead author Joseph Poore from Oxford and his LCA number cruncher T. Nemecek.

As a sidebar, I had meant to write a much more detailed deconstruction of this meta-analysis going over the methodology and conclusions, but simply never got around to it. Needless to say, there were many flaws in their methodology resulting in many papers being excluded. For example, all the papers by Teague, Rowntree, etc since these papers either came too late or weren’t full LCA’s, meaning they looked at only a phase rather than the entire life cycle analysis of a certain food’s production. The other obvious problem with such a methodology is that it looks at averages with ranges of variability  primarily from prevailing industrial systems. Many of the LCA’s were also in Europe. The reality is that the systems that are really going to solve many of our problems are regenerative systems of food production that currently are outliers and thus only more recently being peer reviewed (e.g. Stanley et al. 2018). Based on averages, the numbers favor prevailing industrial systems. Optimizing industrial production still results in massive soil degradation, pesticide use, and blue water use. The best worst case scenario based upon using industrial Ag differently is still not a very good scenario to promote or follow.

Anyway with that aside aside, again the clipped Guardian stat again goes to super simplistic understanding of land use, and doesn’t understand basic nutrition in regards to proteins. As previously mentioned, most land isn’t arable but can be used for grazing. Livestock can also be integrated into arable land. When this is well done, this can reduce or eliminate tillage, eliminate synthetic fertilizer inputs, and reduce or eliminate pesticide use while increasing farmers profits from stacked enterprises and reduced input costs.  As for proteins, using the same data, shows that the 38% of grams of protein translates to much higher percentages (65 to 80%) of specific essential amino acids (e.g leucine, isoleucine, lysine, etc)  provided by animal sources  since meats are nutrient dense complete proteins. Whereas  the vast majority of plants sources are not nutrient dense and not complete sources of protein. Moreover, in the calorie analysis, a large percent of those calories are coming from vegetable oils. Is all the land being used for this calorie production being properly attributed to this oil production?

Whether water footprint numbers, relative risks, land use figures, feed efficiencies or “grams of proteins”, statistics used to promote white hat biases, technologies, food products and or food religions tend to oversimplify what exactly such numbers mean. Too many writers and others in media simply regurgitate such numbers without the slightest bit of critical analysis.  As I’ve previously shown with water footprint numbers, such statistics can also be quite deceiving when taken out of context. For many who aren’t terribly scientifically literate, one senses that citing out of context numbers/stats give such people a sense of intellectual superiority. Though the reality is that reliance upon statistics, with little to no understanding of how such statistics are or were derived,  sadly just further reveals the degree of disconnection that so many, including George Monbiot, have from where and how their food is actually grown, raised or caught. If George ever wants to come to California, and take a tour of regenerative ranches in my state restoring ecosystems for greater biodiversity, he’s more than welcome to reach out to me and I’ll make all the necessary arrangements.

We need to regenerate land through building healthy soil ecosystems. Since without healthy soils, it really won’t matter what dietary pattern anyone follows, because we’re all going to be screwed. At the rate we’re losing top soil, according to the UN , we only have 40 to 60 harvests left.