Livestock, Externalities, and the Environment

The Wall Street Journal published a piece  today that I wrote dealing with externalities in livestock production.  I didn't choose the title - my argument isn't that livestock production doesn't have environmental impacts, rather I question the relative size of the impacts and discuss the best way to handle those impacts.

A few snippets:

“That the price of meat is too low might come as news to food consumers who, according to data from the Bureau of Labor Statistics, paid 14% higher prices for ground beef this June than they did in June 2013 and 29% more than two years ago. Recent droughts and high corn prices—due in part to Washington’s support for ethanol—are largely to blame. It is unclear how high prices must rise to overcome the view that meat is “too cheap.” Some industry critics have even called for new “meat taxes” to discourage consumption.”

and

“The Environmental Protection Agency estimates that U.S. agriculture, including livestock production, accounts for only about 8% of total greenhouse-gas emissions in the country. Livestock in the U.S. have lower greenhouse-gas footprints than in other parts of the world. This is partly because American producers generally use higher-quality feeds, higher-yielding breeds, and more productivity-enhancing technologies such as probiotics, vaccines and growth hormones. Future improvements in feed and animal genetics could further reduce animal-agriculture’s impact. As economists have shown, one should not underestimate the ability of innovation, markets, the courts and private negotiation to resolve the adverse effects of externalities.

Moreover, the concept of externalities when applied to food is nebulous. At a recent Institute of Medicine meeting I attended, a room full of Ph.D.s struggled to understand exactly what to measure.”

and

“Let us also not gloss over what is beef’s most obvious benefit: Livestock take inedible grasses and untasty grains and convert them into a protein-packed food most humans love to eat. We may be able to reduce our impact on the environment by eating less meat, but we can also do the same by using science to make livestock more productive and environmentally friendly.”

For more on that last point, see my previous post.

The piece was in part motivated by the fact that social commentators’ accounts of externalities often reflect a shallow understanding of complexity of the subject.  The economists A.H. Barnett and Bruce Yandle accurately discerned the fact that, “economists unwittingly developed a weapon of mass destruction that, in the hands of journalists and popular policy analysts, at times corroded almost to the point of uselessness the beneficial theory of markets and competition.”  As a participant in one of the CDC-IOM planning workshop on “Exploring the True Costs of Food”, I have witnessed the disconnect that often exists between public health advocates and economists on the nature and role of externalities (I discuss some that disconnect and the complexity of externalities in this article published in Agricultural and Resource Economics Review).  Often, factors that are argued to be externalities are simply zero-sum transfers (as is the case for health care costs paid by public insurance programs like Medicaid), have effects that are actually internalized in other market prices (such as the risk of injury to workers in meat packing plants), or are not externalities at all. 

If the issue is that livestock are consuming "too much" water and that water isn't appropriately priced, the key is to think about how to develop water rights and markets so that the price of water reflects its relative scarcity.  But, it should also be clear - given the correlation between drought and beef prices - that a lot of the water use is factored into the price of beef.

That there are externalities in beef production is hardly news.  The much more difficult question is how to address them.  Technological progress is a key solution.  Research shows that the carbon footprint of beef production fell 16% from 1977 to 2007, with much of that reduction resulting from responsible use of technologies.  Many consumers are averse to these externality-reducing practices and technologies, but more “natural” production systems are often associated with lower productivity, greater water and land use, and higher carbon footprints.  

Hormones in Soybeans and Beef

About 90% feedlot cattle in the US are administered some type of growth hormone to promote growth.  Use of the hormones convey economic benefits to consumers (lower prices) and a host of environmental benefits (more meat using less land, less water, less C02).  The biggest drawback, from my perspective, is the evidence that use of such hormones reduces the eating quality of steaks, particularly by reducing tenderness.  

While reduced tenderness might be a reason to eschew hormones, food safety isn't.  Some people are worried about the health effects of these hormones, but such concerns do not mesh well with the scientific literature, and the concerns tend to ignore relative risk.  Specifically, there are much higher levels of naturally occurring hormone-like substances in many foods we eat.

As a result, there have been many attempts to communicate this information to the public.  Examples of such discussions appear at BeefMyths.org, US Meat Export Federation, the NCBA, and extension facts sheets from Michigan State University, University of Nebraska,University of Georgia, and many others.  

A common approach is to compare the extra amount of estrogen in a serving of beef from an animal that has received a hormone implant to one that hasn't, and then compare that to estrogen-like substances in other foods like soybean oil (it is a comparison I've made myself in a study on the effectiveness of such communication), cabbage, peas, and potatoes.

After making this comparison in a talk a couple weeks ago, an audience member gently questioned my numbers on soybean oil.  While it is true that soybeans have high levels of isoflavones, which acts like estrogen in humans, it turns out that these compounds are not in soybean oil.  

Here is a publication from the USDA Ag Research Service showing the isoflavone content of a long list of foods.  As you can see, soybeans have quite a bit, but if you'll look down on page 38, you'll find soybean oil listed in a table titled "List of Foods Containing Zero Values for Isoflavones."  This website neatly summarizes the USDA data.

So, where does that leave us.  First, those that have used this comparison should try to correct the record (as I'm doing here).  If we are arguing that the public should make decisions on "the facts," we darn well better get our facts straight.  Second, the relative hormone comparison remains useful (though only marginally persuasive with most consumers), but one needs to drop soybean oil and use other soy products instead.  An Iowa State University Fact's Sheet by Dan Loy helps make the proper comparisons.

Here is a key screenshot

How should food policy issues be decided?

Not only is it the case that people are likely to differ in their opinions about the desirability of mandatory GMO labeling or soda taxes, but they are also likely to differ in how they think such issues should be decided.  

A while back, I ran across this paper by Gaskell and colleagues published in Science.  They sought to categorized citizens in terms of their attitudes about how technology should be governed by asking two questions relating to whether decisions about technology should be made by 1) experts vs. average Americans and 2) moral and ethical issues vs. scientific evidence on benefits/costs.  

In the latest edition of the Food Demand Survey (FooDS), I applied these questions to five food policy issues.  Unlike Gaskell's work I also allowed respondents to have different answers for different issues.  

The first question asked: “Decisions about food policy should be based mainly on the views and advice of experts OR decisions about food policy should be based mainly on the views of the average American.”  The second question asked: “Decisions about food policy should be based mainly on the moral and ethical issues involved OR decisions about food policy should be based mainly on the scientific evidence of risk and benefit.”  Then five food policy issues were listed in random order: labeling of genetically modified food, use of growth hormones, legality of selling raw, unpasteurized milk, use of the term "natural" on packaging, and the tax charged on sugar sodas.

Here's what we found.

More than 70% of respondents wanted policy decisions related to GMO labeling and use of growth hormones to be based on expert advice rather than the views of the average American. I find that result rather striking in light of the fact that opinion polls show large numbers of people saying they want GMO labeling.  Here, we see that a large majority thinks this sort of issue should NOT be decided by the views of the average American.  That would seem to imply that folks do not think GMO labeling should be settled by ballot initiative.  

In stark contrast to the other food policy issues, almost 70% wanted decisions about soda taxes to be based on the views of the average American rather than the "elites".  

Recall that we also asked about whether decisions should be based on morals and ethics or based on scientific evidence on risk and benefit.

For three issues, milk pasteurization, hormones, and GMO labeling, the majority thought decisions should be based on science.  There was a split on natural labeling.  For soda taxes, the majority thought moral issues should be the deciding factor.

As with the prior research, we used the answers to categorized people into one of four categories for each of the five food policy issues.  “Scientific elitists” wanted policy decisions made by experts on the basis of scientific evidence, “moral elitists” wanted policy decisions based made by experts on the basis of moral issues, “scientific populists” wanted the average American to make decisions on the basis of scientific evidence, and “moral populists” wanted the average American to make decisions on the basis of moral issues.

A plurality of respondents were "scientific elitists" for GMO labeling, use of growth hormones, and legality of selling raw milk.  The same was true for use of the term "natural" on labeling, but there was a larger share of "moral elitists" in regard to this issue than for others.  Finally, for soda taxes, "moral populists" described the largest share of respondents.

A natural question is whether these categories explain people's attitudes about the food policies.  Gaskell et al. showed that "scientific elitists" in regard to general technology were the majority citizen type in their surveys and this type had more favorable attitudes toward biotechnology and nanotechnology than other consumer types - particularly moral populists.

I find something similar here as well.  Take for example, the 4-category breakdown on GMO labeling.  I find that "scientific elitists" on GMO labeling express the lowest level of concern about eating GMOs (an average score of 3.06 on a 1 to 5 scale of concern), whereas "moral elitists" and "scientific populists" had scores of 3.41 and 3.43.  Moral populists averaged 3.34.  There also seems to be a political dimension to people's views about how these food policy issues should be decided.  For example, scientific elitists and scientific populists were slightly more conservative (about 3.05 on a 1 to 5 scale of liberal to conservative) than were those who focus more on moral/ethical issues (score of about 2.9 on the scale).  Those identifying with the Democratic party were more heavily represented in the "moral elitist" category than they were in other categories.  

Critique of modern agriculture in song

One of our gradudate students sent me a link to this music video performed by a band, Wookiefoot, he recently saw in concert.

If you couldn't follow along, here's a sample of some of the lyrics

Then McDonalds got a farm GM GM GMO
And on that farm he had a chicken and a cow
But they do now you really don’t wanna know
Why must we label it organic
When that’s the way we’ve been growing it for ten thousand years

I don't suppose it would do any good to mention what organic really means or to show what agriculture really looked like 10,000 years ago?  

Does information on relative risks change concerns about growth hormones?

Consumers often express concern about the use of growth promotants in animal agriculture.  In the beef industry, various growth hormones are administered to cattle to improve and speed the rate of growth (and some would say, improve the sustainability of beef production).  Upwards of 90% or more of feedlot cattle in large feedyards are given hormone implants.

Some consumers are fearful about the safety effects.   For example, the EU has banned imports of hormone-treated cattle from the US for over 20 years (a policy which probably has more to do with protectionism than actual safety concerns).  Other people have argued that these are the cause of decreasing puberty age of girls (which the data doesn't support).

As a result, many in the beef industry have have tried to communicate the fact that the risks from hormones are small to non-existent, and are much smaller than the risks from hormones in everyday foods.  The normal comparison is between how much estrogen is in a hamburger from an implanted steer or heifer vs. the amount of estrogen in other foods like soybean oil or cabbage.  Examples of such discussions appear at BeefMyths.org, US Meat Export Federation, the NCBA, and extension facts sheets from Michigan State University, University of Nebraska, University of Georgia, and many others.

Circulating on the web a while back were some discussions of using some visual strategies to communicate the relative risks from estrogen used in cattle implants.  For example, here is one blog discussing the use of M&Ms to convey the risks.  

The question I wanted to know is whether any of these sorts of communications actually has any impact on the people for whom it is intended.  

In the most recent issue of my monthly Food Demand Survey (FooDS), we sought to address this issue.  1,017 respondents were randomly allocated to one of three information groups or treatments.  In the first no-info group, respondents were simply told, “About 90% of feedlot cattle are given added growth hormones to improve the rate of growth.” And then, respondents were asked, “How concerned are you about the use of growth hormones in beef production?”  

For the second group text-only group, written text was added to convey relative risks of hormone use.  Prior to being asked level of concern, subjects were told, “About 90% of feedlot cattle are given added growth hormones to improve the rate of growth.  The added hormones add about 3 extra nanograms (a billionth of a gram) to a 3 oz serving of beef.  For comparison purposes, the amount of estrogen that naturally occurs in 3 oz of the following foods is: potatoes (225 nanograms), peas (340 nanograms), cabbage (2,000 nanograms), soybean oil (170,000 nanograms).”  

Finally, the third visual+text group was given the same written text but was also shown the above visual illustration using M&Ms allocated to different jars.  

Participants in all three groups answered with their level of concern on a five-point scale (1 = very unconcerned; 5=very concerned).

Information on relative risks caused a small but statistically significant reduction in the level of concern.  The mean levels of concern, on the 5-point scale, were 3.93, 3.71, and 3.66 for the no-info, text-only, and text+visual information groups.  

Without any information on relative risks, over 71% of respondents indicated that they were either concerned or very concerned.  Textual information reduced that frequency to 66%, and visual+text information further reduced the percentage of concerned respondents to 63.6%.