Sunday, November 01, 2015

Trends in farmed animal life-years per kg and per human in the United States

Summary: Selective breeding, drugs, and altered diets have greatly increased the quantity of milk, meat, and eggs produced per year of farmed animal life for multiple species, creating side effects that lowered the quality of life of farmed animals, and increasing consumption through lower prices. In the United States it appears that for some agricultural industries productivity increases since 1950 might have reduced farmed animal-years enough to outweigh the effect of falling prices. These include dairy, beef, and eggs. However, total animal-years of chickens raised for meat, the most populous farmed land animal, increased dramatically in total and per capita, despite a severalfold reduction in chicken-years per kg of meat sold. Increases in production efficiency may reduce demand for farmed animal-years in some mature developed country markets, while increasing demand in larger emerging markets. Further analysis using detailed income and price elasticity information, as well as welfare effects of overbreeding, could estimate net effects of technological change on animal welfare.


Selective breeding in dairy cattle
Wikipedia reports that:

The United States dairy herd produced 83.9 billion kg (185 billion lbs) of milk in 2007, up from 52.6 billion kg (116 billion lbs) in 1950,[2][3] yet there only about 9 million cows on U.S. dairy farms—about 13 million fewer than there were in 1950.[3]

In other words, the average quantity of milk produced per cow has more than quadrupled, and this has led to a decline in the number of cows in U.S. dairy farms of more than 60%. The U.S. population increased from 152.3 million to 301.2 million over that period, so per human consumption of milk declined despite the fall in prices.

In substantial part this change has occurred through intense selective breeding, including artificial insemination,  IVF,  pedigrees, genomic prediction, and other methods. Moreover, the example of existing individuals, as well as predictions from genomics data, indicate the process could continue to even more extreme levels:

Jarrod Kollwelter has exceeded a rolling herd average of 40,000 pounds – one of the highest in the nation. Yet, the young and ambitious dairyman is striving for more milk, and he strongly believes that genetics is the golden ticket to taking his herd to the next level.
“I really believe by the year 2020 we should be able to push 50,000 pounds of milk,” Kollwelter stated at the 2013 Alltech Dairy School. “And genetics is what will get me there.” 
Kollwelter milks 210 cows at his JC-Kow Farms near Whitewater, Wisconsin. The third-generation farmer with a passion for genetics and impeccable cow sense has expanded the herd from within since returning home from college in 2002.
At the same time, his herd average has been on a steady climb. Presently, it is right around 40,280 pounds of milk with no signs of slowing down...

His highest producing cow, a Planet that made 56,040 pounds of milk in her third lactation, “bred back on one shot every time.” 
It is thus very reasonable to predict that successive generations of breeding using existing techniques will be able to achieve further reductions of several-fold in the number of cow-years per kg of dairy products, and that this could even lead to further declines in the dairy cow population.

However, the same genetic changes and hormones that have reduced the number of cows farmed in dairies also increase stress on the body, especially the udder, leading to high rates of painful mastitis. PETA:
After their calves have been taken away from them, mother cows are hooked up, two or more times a day, to milking machines. Their reproductive systems are exploited through genetic selection, despite the negative effects on their health. Artificial insemination, milking regimens, and sometimes drugs are used to force them to produce even more milk—the average cow today produces more than four times as much milk as cows did in 1950. 
Cows may be dosed with recombinant bovine growth hormone (rBGH), which contributes to an increased incidence of mastitis, a painful inflammation of the udder. (In the U.S., rBGH is still used, but it has been banned in Canada and the European Union because of concerns about human health and animal welfare.) According to the U.S. Department of Agriculture, 16.5 percent of cows used for their milk suffer from mastitis, which is one of the leading causes of death in adult cows in the dairy industry.
I would guess that if one thinks the lives of dairy cattle to be on average worse than nonexistence without mastitis, then the very large fall in the size of the dairy cow population, especially per human dairy consumer, could mean that rBGH, antibiotics, and selective breeding have led to net cattle welfare improvements in the United States, although advances may have increased consumption elsewhere.

Cattle killed for beef
In animals farmed and slaughtered for meat, the number of animal-years per kg is affected not only by the amount per animal, but also by the lifetime required to reach that size. Michael Pollan writes about the decline in lifespan before slaughter in cows killed for meat, especially in factory farm feedlot operations:
So if this system is so ideal, why is it that my cow hasn't tasted a blade of grass since October? Speed, in a word. Cows raised on grass simply take longer to reach slaughter weight than cows raised on a richer diet, and the modern meat industry has devoted itself to shortening a beef calf's allotted time on earth. ''In my grandfather's day, steers were 4 or 5 years old at slaughter,'' explained Rich Blair, who, at 45, is the younger of the brothers by four years. ''In the 50's, when my father was ranching, it was 2 or 3. Now we get there at 14 to 16 months.'' Fast food indeed. What gets a beef calf from 80 to 1,200 pounds in 14 months are enormous quantities of corn, protein supplements -- and drugs, including growth hormones.
This shift has also involved extensive genetic change using advanced reproductive technologies and statistics. And this has resulted in a fall in the size of the beef cattle herd relative to beef consumption:
In 1976 per capita beef consumption peaked in the United States at 91.5 pounds a year. It has since fallen more than 40 percent. Last year Americans ate on average 54 pounds of beef each, about the same amount as a century ago. Instead we eat twice as much chicken as we did in 1976 and nearly six times as much as a century ago. It’s cheaper and supposedly better for our hearts. We slaughter more than eight billion chickens a year now in the U.S., compared with some 33 million cattle...

In 2013 the U.S. produced almost the same amount of beef as it did in 1976, about 13 million tons. It achieved this while slaughtering 10 million fewer cattle, from a herd that was almost 40 million head smaller. The average slaughter animal packs 23 percent more meat these days than in 1976. To the people at Cactus Feeders, that’s a technological success story—one that meat producers will need to expand on as global demand for meat keeps rising.
However, the productivity gains were smaller than in dairy and the overall size of the herd did not shrink, with the total herd size similar in 2014 and 1950.

Egg-laying chickens
The FAO reports that the number of eggs per chicken increased from 170 in 1925 to 325 in 2006. 1900 production was 83 eggs per hen per year. The size of eggs has also increased.

The 1950 USDA chicken and eggs report gives an inventory of 481 million non-broiler chickens producing 56 billion eggs. In 1965 the figures were 65 billion eggs, 216.5 per layer, with instantaneous inventory of 367 million chickens. In 2015  the number of layer chickens varied between 330 million and 360 million, with total production of 100 billion eggs for sale, 87 billion for sale (others were used to produce new layer and broiler chicks, overwhelmingly the latter). Thus the total population of chickens producing eggs for sale (or such chickens) was a little over 300 million.

Since 1950 layer chicken-years per egg fell by almost 60% and the total layer chicken population in the U.S. fell. It also appears that breeding for egg production may have produced less severe welfare problems  than the mastitis in dairy cattle, or leg problems for broiler chickens: PETA's description of welfare issues in the egg industry focuses on the horrendous conditions of battery cages, disease, overcrowding, mutilation, and slaughter rather than effects of overbreeding.

Chickens raised for meat
'Broiler chickens' bred for meat production have seen a steep decline in chicken-years per kg, but per capita demand in the United States skyrocketed, as consumers substituted away from other higher cost meat products with slower price improvements and responded to concerns about health effects of red meat.

Vox, drawing on(Zuidhof et al) offers this image showing birds from genetic lines reflecting the state of the art in different years but fed the same diet and at the same age:

Giant chickens with dates

The image exaggerates differences at slaughter, because the measurements were taken before the older breeds had time to mature, but convey the magnitude of change. The excessive size causes lameness, leg infections, and heart problems.

Change in U.S. broilers (source: U.S. National Chicken Council):

1925 to Present
YearMarket Age

average days
Market Weight

pounds, liveweight
Days Per Pound

average days
Feed to Meat Gain

pounds of feed to one pound of broiler, liveweight
Mortality

percent
19251122.544.84.718
1935982.8634.34.414
1940852.8929.4412
1945843.0327.7410
1950703.0822.738
1955703.0722.837
1960633.3518.82.56
1965633.4818.12.46
1970563.6215.52.255
1975563.7614.92.15
1980533.9313.52.055
1985494.1911.725
1990484.371125
1995474.6710.11.955
2000475.039.31.955
2005485.378.91.954
2006485.478.81.965
2007485.518.71.954.5
2008485.588.61.934.3
2009475.598.41.924.1
2010475.78.21.924
2011475.828.11.923.9
2012475.957.91.93.7
2013476.017.71.883.7
2014476.127.51.894.3
Mid-2015486.27.71.915.2
The reduction in days per pound from 44.8 in 1925, to 22.7 in 1950, to 7.7 today is slightly supplemented by the decline in mortality rates (which reflected additional years of suffering without increases in meat sales). So broiler chicken-years per kg have declined fairly dramatically, although these changes have directly increased physiological problems related to size.

Unfortunately, broiler populations grew much faster: in the U.S. over 8.5 billion broilers were killed in 2013, up from 4.48 billion in 1985, and  487 million in the 1950 USDA chicken and eggs report. Even in terms of per human animal-years, that represents a tripling since 1950, as the broiler industry initially took off. Between 1985 and 2013 pounds of broiler meat consumed increased from 18.9 billion to 50.7 billion, while the human population increased from 237.9 million to 316.5 million, doubling per capita consumption. Chicken Council numbers:


1/ Includes beef, pork, veal, and mutton/lamb, but excludes edible offals.
2/ Estimated by USDA
Note: All poultry and livestock products are on a retail weight basis, except “other chicken” and “turkey” which are reported by USDA on a carcass-weight basis.
Fish/shellfish is reported by The National Marine Fisheries Service on an edible weight basis.
Source: USDA
Last updated July 10, 2015

As the lowest-cost meat product with the fastest price declines, chicken gained market share at the expense of other animal product industries.

International consumption, income and substitution effects
The above data are drawn from the United States, a relatively rich country where humans have been able to afford large amounts of meat for many decades, and consumption may be closer to saturated. However, emerging economies are collectively far more populous and currently have much lower meat consumption. Technological change which reduces meat prices may drive larger increases in demand there.

To estimate how much technological change and falling prices contributed to consumption of different animal products, one would have to look at income elasticity of consumption, price elasticity and substitution between different animal products. I invite interested readers to try this, but this post will simply share the data I encountered.

5 comments:

Pablo said...

>I would guess that if one thinks the lives of dairy cattle to be on average worse than nonexistence without mastitis, then the very large fall in the size of the dairy cow population, especially per human dairy consumer, could mean that rBGH, antibiotics, and selective breeding have led to net cattle welfare improvements in the United States, although advances may have increased consumption elsewhere.

On the other hand, if the lives of dairy cattle are assumed to be on average better than nonexistence even with mastitis, as F. Bailey Norwood in fact assumes (p. 229), then the opposite conclusion follows. Brian Tomasik believes that Norwood's estimates are very optimistic; unfortunately these are, to my knowledge, the only estimates we have.

Carl said...

Pablo, a good project for factory farming prioritization might involve doing such estimates with a larger sample size than n=1, better definitions, and with good access to data for the raters (who could be drawn from philosophers, biologists, veterinarians, and so on).

Douglas Knight said...

I am not sure you that you infer anything from it, but the claims in your second blockquote are incorrect. (The Atlantic misunderstood its poorly labeled sources.) The numbers it quotes are milk production per year, not lifetime. In fact, I believe that the lifetime production has not changed, the 4x productivity compensated by 1/4 lifespan, mainly due to aggressive culling. Since most of your discussion is about years, not lives, the correct figures are more relevant.

Incidentally, I believe that the 4x productivity is due to a combination of 2x productivity while lactating and 2x time spent lactating.

Carl said...

Thanks Douglas, I removed the quote.

Unknown said...

I'm late to this discussion, but Dr. Sara Shields, an animal welfare scientist at the Humane Society, has also assigned some welfare scores on Norwood's scale. In the case of dairy cattle, Shields gives a score of 0, which is supposed to mean "The positive emotions experienced by the animal is equal to negative emotions. The animal is indifferent to living."

See this excerpt from Veganomics by Nick Cooney, reproduced here (with permission) by Matt Ball:

http://www.mattball.org/2014/07/part-1-analyzing-numbers-to-optimize.html

This is the scale that is supposed to be used:

https://snag.gy/abgcyj.jpg

I should note that I've sent some time trying to understand how Norwood and Shields arrived at these scores, and have reached the conclusion that both of their scores have some major problems.