Blog

Twitter conversations about GMOs

Last week, an organization called Right Relevance, put out a fascinating post analyzing Twitter interactions surrounding the topic of GMOs during a single month - January 2017.  I don't claim to fully understand all the methods they used or precisely how to interpret the figures they generated, but here's one of their conclusions:

The retweets-only graph (Fig 2) is even more stark in bringing out the partisanship. It visualizes the echo-chamber like nature of the partisan groups. Also, it shows higher diversity and broader participation on the anti-GMO side.

The go on to document and rank popular themes, topics, and individuals.  I was a bit curious about the graphs, and even though I didn't recall tweeting much about GMOs in January of 2017, I though I saw my name in tiny font next to NYT Science in the above graph, so I emailed the author of the post asking for a higher resolution figure.  Instead, they sent me the following two graphs focused specifically on my Twitter account (the second one I believe is only based on re-tweets). 

I suppose I shouldn't be at all surprised to recognize most of the names in these figures since they're the same people I'm interacting with on Twitter.  Still, there are many names I don't recognize but who are apparently in my "network".  I'm not sure whether I should be frustrated that my Twitter network on this topic isn't bigger and more diverse or just be thankful for the network I have.  It would also be interesting to see these same figures at different points in time.  From personal experience, I can tell you that when I've had articles on GMOs in the New York Times or Wall Street Journal, I get a lot of people tweeting at me that have widely opposing views.  

The intractability of the soda tax debate

This recent article claims that the new soda taxes in Philadelphia are causing a larger than expected drop in soda sales.  

Did that piece of news change your mind on whether the soda tax is a good or bad idea?  My guess is: probably not.  As humans, we're adept at finding ways of confirming our prior beliefs and positions.  That is, we suffer from various forms of confirmation bias.  

Let's take the above story at face value: after the tax was implemented, we find a large reduction in soda consumption.  What are the expected reactions by the competing camps? (note: by definition, a large reduction in soda consumption implies that the receipts from the tax are smaller than expected as people are avoiding the tax by buying less soda).

The pro-tax folks would say:

  • "Ah-ah!  This simple solution has a big public health benefit.  It got people to stop consuming all those useless, empty calories, and now we'll finally make headway on obesity and diabetes.”
  • “This big drop in consumption was accomplished without costing consumers much.  In fact their expenditures on soda fell!  Now they have more money to spend on healthier items.”  

The anti-tax folks would say:

  • “We told you this tax policy would cost jobs.  Nobody is buying soda anymore, and people will have to be laid off at the beverage manufacturing plants and the beverage distributors.”
  • “You promised that the tax would fund public education but there aren’t enough new tax receipts to fund any new programs or to give teachers meaningful raises.”
  • “People may have stopped buying soda, but look now they’re buying more [insert the untaxed, unhealthy food of your choice here].  

Now, instead imagine the opposite case was observed.  Suppose that after the tax was implemented, we find no (or a small) reduction in soda consumption.  What are the possible reactions by the competing camps? (note: by definition, a small reduction in soda consumption would imply that the receipts from the tax are higher than expected – the government is raking in money as people are still buying soda and paying the tax).

The pro-tax folks would say:

  • “Look at all the new money we’ve raised to finally get to work on [insert your favorite public program or cause here].”
  • “We're finally making "Big Soda" pay for all the costs they've been imposing on society.”
  • "The soda tax is just one small part of an overall plan to reduce obesity and improve public health." 

The anti-tax folks would say:

  • “This policy created a bureaucratic agency to oversee the tax, increased the size of government, and look, it didn’t have any impact on obesity or public health as promised.”
  • “We told you this was a regressive tax.  The majority of the new tax dollars being generated are being paid by lower income households.”
     

This sort of conundrum shows that it is hard to have an intellectually honest debate about the evidence.  It also suggests that policy advocacy (or policy opposition) is often more about competing values or philosophies than it is about empirical evidence (after all, both sides claim to want evidence-based decision making). In general, I think it is troubling if someone can't answer the question: "What evidence would it take to convince you that you were wrong?"

As more locations - from Seattle to Santa Fe - are considering the adoption of soda taxes, it would be useful if folks stated, before adoption takes place, what outcomes would or would not support their initial opposition or advocacy to the tax.  Because right now, any post-tax outcome can be interpreted as evidence in favor of either position. Or, just be honest, and say that opposition or advocacy for the tax is not an evidence-based position, but rather one based on some underlying philosophy or set of values.

As for me, I'll admit that much of my opposition to soda taxes is indeed value-based based.  I tend to favor freedom of choice and limited government, and I haven't been convinced by the market-failure arguments that would justify the tax.  Now, let me put that to the side and say that most of my writing on the subject has argued that, empirically, soda taxes are unlikely to have much effect on obesity/diabetes rates.  As a result, I see the policy as an ineffective means to achieve the desired end (by the way if we want to fund more public education, there are likely more efficient was of doing that than taxing soda).  Moreover, if it is true that the tax doesn't much change consumption, it implies consumer demand is relatively inelastic, which also implies that the tax burden primarily falls on the consumer rather than the producer (note: economic theory indicates that it doesn't matter whether the tax is technically imposed on the producer or the consumer - it is the underlying elasticities of supply and demand that determine who actually bears the burden of the tax). So, if the claim in the above article is true and the soda tax has a "large" effect on soda consumption, that would undermine many of my empirical arguments against the soda tax.  

But, I'm not sure that evidence showing that a soda tax had a large reduction in soda consumption would turn me into a tax advocate.  After all, if you showed me that a tax on broccoli caused a large reduction in broccoli consumption, I wouldn't suddenly become a broccoli-tax advocate.  Rather, I think the kind of evidence I'd find more persuasive, if one wanted to substantively more me away from an anti-soda tax position, is evidence that soda consumption causes an externality (and "no" the presence of Medicare/Medicaid is not evidence of an efficiency-reducing externality) or that, in this particular case, there were substantively perverse information asymmetries (although the appropriate policy here is probably information provision not a tax), or behavioral biases that citizens themselves want "corrected" via government action (just because a soda tax passes with more than 50% of the vote doesn't really "count" as evidence here because many of the people who vote in favor of the tax don't consume much taxed soda).  

Rising Input Costs and Farmers' Responses

Low commodity prices have prompted farmers to be ever more cognizant of their costs.  I've run across a couple interesting articles of late that touch on this issue.  The first is a piece by David Widmar, who discusses the rising cost of seed for farmers.  He writes the following about corn seed prices and provides the associated figure:

Since 2002 seed expense has increased rapidly. From 2006 to 2014 alone, the seed expense nearly doubled in real dollars; increasing from $51 per acre in 2006 to $101 per acre in 2014 (in nominal dollars, the increase was from $44 to $101).

He concludes:

this environment seems ripe for a seed manufacture to cut prices as a strategy for gaining market share.

While farmers are waiting on incumbent firms to cut prices or for new low-cost entrants, they can act too.  This article in the Wall Street Journal discusses how some farmers are turning to online outlets to source lower priced inputs such as fertilizers and herbicides.  Here's one excerpt:

Online sellers, including some wholesale distributors and national farm retailers, often offer generic versions of popular pesticides that are cheaper than the branded counterparts frequently sold by co-ops. FBN says it also can offer products at a discount because it lacks expenses associated with brick-and-mortar facilities and is able to get better deals from manufacturers because of its national scale.

“It’s infuriating knowing how deeply we’ve been gouged,” said Nebraskan farmer Clay Govier, who said he saved more than $12,000 on herbicides for his 3,000-acre corn and soybean farm by checking prices on FBN’s network.

Measuring Beef Demand

There has been a lot of negative publicity about the health and environmental impacts of meat eating lately.  Has this reduced consumers' demand for beef?  Commodity organizations like the Beef Board run ads like "Beef It's What's for Dinner."  Have these ads increased beef demand?  To answer these sorts of questions, one needs a measure of consumer demand for beef.  In my FooDS project, I try to measure this by using consumers' willingness-to-pay for meat cuts over time.  But, there are other ways.

I just ran across this fascinating report Glynn Tonsor and Ted Scroeder wrote on beef demand.  At the onset, they explain their overall approach.

One way to synthesize beef demand is through construction of an index that measures and tracks changes in demand over time. An index is appealing because it provides an easy to understand, single-measure indicator of beef demand change over time. A demand index can be created by inferring the price one would expect to observe if demand was unchanged with that experienced in a base year (Tonsor, 2010). The “inferred” constant-demand price is compared to the beef price actually transpiring in the marketplace to indicate changes in underlying demand. If the realized beef price is higher (lower) than what is expected if demand were constant, economists say demand has increased (decreased) by the percentage difference detected. Applying this approach to publically available annual USDA aggregate beef disappearance and BLS retail price data provides information such as contained in Figure 1 indicating notable demand growth between 2010 and 2015 based upon existing indices currently maintained at Kansas State University.

They then show the beef demand index that Glynn has been updating for several years now based on aggregate USDA data.

In their report, Tonsor and Schroeder show, however, that measures of beef demand depend greatly on: 1) the data source being used, 2) the cut of beef in question, and 3) consumers' region of residence.  For example, here is a different beef demand index based on data from restaurants (or the "food service sector") segmented into different types of beef.  You'll notice the pattern of results below differs quite a bit from the aggregate measure above.  And, whereas demand for steak fell during the recession, demand for ground beef rose.

Another interesting result from their study is that the commonly used retail beef price series reported by the Bureau of Labor Statistics doesn't always mesh well with what we learn from from retail scanner data (in their case, data from the compiled by the company IRI).  Not only are BLS prices a biased estimate of scanner data prices, the bias isn’t constant over time.  In the report, Tonsor and Schroeder speculate a bit on why this is the case.  

In the near future, Glynn and I aim to compare my demand measures from FooDS with these demand measures. 

Banning Soda Purchases Using Food Stamps - Good idea or bad?

According to Politico:

The House Agriculture Committee this morning is delving into one of the most controversial topics surrounding the Supplemental Nutrition Assistance Program: whether to limit what the more than 40 million SNAP recipients can buy with their benefits. Banning SNAP recipients from being able to buy, say, sugary drinks has gotten some traction in certain public health and far-right circles, but it looks like the committee’s hearing will be decidedly open-minded on the debate.

I've written about this policy proposal several times in the past.  It's an example of good intentions getting ahead of good evidence.  Do SNAP (aka "food stamp") participants generally drink more soda than non-SNAP participants?  Yes.  Is excess soda consumption likely to lead to health problems?  Yes.  But, will banning soda purchases using SNAP funds reduce soda consumption.  Probably not much.  

In fact, I just received word that the journal Food Policy will publish a paper I wrote with my former Ph.D. student, Amanda Weaver, on this very topic.  First is the logical (or theoretical) argument:

In public health discussions, however, the conceptual arguments related to the Southworth hypothesis have received scant attention (see Alston et al., 2009, for an exception). A soda consuming SNAP recipient who spends more money on food and drink than they receive in SNAP benefits can achieve the same consumption bundle regardless of whether SNAP dollars are prohibited from being used on soda by rearranging which items are bought with SNAP dollars and which are bought with other income. Thus, an extension of the Southworth hypothesis to this case would predict little or no effect of a soda restriction as long as the difference in total food spending and SNAP benefits does not exceed spending on sugar-sweetened beverages.

If that wasn't transparent, consider the example I gave in this paper I wrote for the International Journal of Obesity:

To illustrate, consider a SNAP recipient who receives $130 in benefits each month and spends another $200 of their own income on food for total spending of $320. Suppose the individual takes one big shopping trip for the month and piles the cart with food, including a case of Coke costing $10. Suppose the cost of all the items in cart comes to $320. SNAP benefits cannot cover the entire amount, but the individual can place a plastic divider on the grocery conveyer belt, put $130 on one side (to be paid for with the SNAP benefits), and put $200 on the other side (to be paid for with cash). Now, suppose there is a ban on buying soda with SNAP. What happens? The individual can simply move the $10 case of Coke from the SNAP side of the barrier to the cash side and replace it with other items worth $10. The end result is the same regardless of whether the SNAP restriction is in place or not: spend $320 and Coke is purchased.

So, in theory, people can "get around" these sorts of SNAP restrictions very easily making the restriction ineffectual.  

Now, back to my Food Policy paper.  Our experiment results show the following: 

As conjectured by H3, for the 65% of participants (78/120) who did not consume soda in T3, soda expenditures were unaffected soda restriction. H4 posited that consumers who had expenditures of more than $2 (including a soda purchase) in T3 would likewise be unaffected by the soda restriction as they moved to T4. However, this hypothesis was rejected (p<0.001). Soda expenditures fell from an average of $1.000 to $0.588, contrary to the theoretical prediction. We find that 58.8% (20/34) of the respondents to which the hypothesis applied behaved as the theory predicted (they did not change soda expenditures); however, the remaining 41.1% (14/34) reduced soda expenditures when moving from T3 to T4.

So, maybe restrictions on soda purchases by SNAP recipients will affect their soda consumption after all.  Here are our thoughts on that:

Previous research has identified heterogeneity in cognitive abilities and in consistency with economic theories (Choi et al., 2014; Frederick, 2005), and future research might seek to explore the extent to which cogntive ability plays a role in the ability of extramarginal consumers to recognze that they can achieve the same consumption bundle despite the soda restriction. In addition, our experiment was a one-shot game. In a field environment, respondents can talk to friends, gain experience, and alter behavior over time as they learn that the same consumption bundle can be achieved despite the restriction. This learing conjecture could be tested in an experimental setting by conducting repeated trials with feedback. It could also be tested using field data (after a policy was passed) by investigating the change in soda purchases for inframarginal buyers over time. Another hypothesis that could explain the anomolous result is that the soda restriction could have non- pecuinary effects, providing information about realtive healthfulness of items or signaling what people “should” be doing. For example, Kaplan, Taylor, and Villas-Boas (2016) found that, following a widely publisized vote to tax sodas, Berkeley California residents reduced soda consumption before the tax was even put into place, illustrating significant information effects surrounding soda consumption policies. Future research could further explore this signaling effect by including a treatment that restricts purchases of food items not generally percieved as unhealthy or by including survey questions about percieved healhfulnes of an item before and after a restriction.

Another thing to keep in mind is that such restrictions may limit people's willingness to participate in SNAP in the first place.  Even in our experimental context, we find that soda restrictions do indeed affect participation as measured by use of the "coupon" or "stamp" (both whether it is used at all and the amount of the coupon used).  

All in all, I think the above discussion shows that despite the intuitive appeal of a simple policy restricting SNAP purchases, the actual consequences are likely to be much more complicated.