An unplanned shock to beef quality supply

In economics, it's tough to separate correlation from causation because the world is a messy place with lots of things changing at the same time.  As a result, empirical economists are always on the lookout for natural experiments, or situations where there was some random, unanticipated "shock" to the market that can help us get closer to an experimental setting, where we know a change in X was not due to a change in Y.  

I was reading through the latest edition of Meatingplace magazine, and was surprised to see a story about an event that provides precisely the sort of unplanned "shock" that we are always looking for. In particular, about eight years, ago, the USDA started using cameras (rather than people) to determine meat quality.  The two main quality grades are Choice (more marbled (or fattier), higher quality) and Select (leaner, lower quality).  

Apparently in June 2017, the USDA issued an update to USDA's camera grading system that "appeared to inaccurately assess the degree of marbling on some carcasses - allegedly grading some Choice that should have been Select." The USDA issued a new update to the cameras in October in 2017 to correct the problem. One analysis, quoted in the article, estimates that about 12,000 cattle were inadvertently graded Choice rather than Select (a 2.4% increase according to the article, if I'm reading it right).

So, we have an unplanned, unanticipated "shock" to the beef quality market that shifted the supply of high quality meat and reduced the supply of lower quality meat.  This is illustrated by the two vertical lines in the figure (the lines are vertical because the supply is fixed in the short-run: you can't take Choice carcass and turn it into a Select one once the animal has been removed from feed).  If demand curve slopes downward, then this unanticipated increase in supply of Choice (and reduction in Select) quantity, should reduce the price premium for Choice over Select.  And in-fact, because the shock to supply is completely exogenous (it had nothing to do with demand but with a camera update), we should be able to use the natural experiment to estimate the slope of the relative demand curve for high quality beef (or the so-called elasticity of demand).  


Here is data from the USDA on the difference in price between Choice and Select beef, or the so-called Choice-to-Select spread, over the time period of interest (in particular, this is the difference in boxed beef cutout values measured in dollars per hundredweight - or cents per pound).    


Just as one would expect, the increase supply of Choice relative to Select led to a reduction in the price premium charged for Choice relative to Select.  Of course, these raw data might be misleading - what if there is a seasonal pattern in which the Choice-to-Select spread falls every year from June to October?  To address this concern, I downloaded the last 10 years of data on the Choice-to-Select spread and found that the observed Choice-to-Select spread from mid June to late October in 2017 was $4.34/cwt lower than would be expected even after controlling for seasonality (month of the year), year, and a time trend.  This works out to about a 31% lower Choice-to-Select spread than would have expected during this time had it not been for the grading camera update (assuming there aren't other confounds I'm not controlling for).  

So, good news, it appears, the demand curves do indeed slope downward.  We can also go further if we take the aforementioned 2.4% change in quantity at face value that came from the Meatingplace article.  The price flexibility of demand (this is roughly the inverse of the elasticity of demand) for Choice (relative to Select) is given by the percent change in price over quantity, or -31%/2.4% = -12.9%.  So for every 1% increase in the quantity of Choice vs. Select supplied, there is a 12.9% reduction in the Choice-to-Select price spread.  

Factors Affecting Beef Demand

Glynn Tonsor, Ted Schroeder, and I recent completed a report for the Cattlemen's Beef Board on the factors influencing beef demand.  

One of the key factors that emerges from the analysis of the USDA price/quantity data is that beef demand appears to have become less sensitive to price-related factors.  In econ-lingo, beef demand has become more inelastic.  Moreover, changes in pork and poultry prices have fairly small impacts on beef demand.

As a result, we focused on several potential non-price demand determinants.  We find that emerging stories about climate change have adversely affected beef demand, but at the same time increased media focus on taste and flavor have more than compensated for those effects, pulling up demand since 2012.  

We also look at trends from the Food Demand Survey (FooDS) and how they relate to consumers' preferences and beliefs.  Here are some graphs on the relationship between a variety of factors and steak demand.

steak demand.JPG

Here is the same but for ground beef demand.


Increases in income clearly increase steak demand, but ground beef is demanded similarly by all income categories.  Some of the biggest determinants of beef demand are "food values".  Here's what we have to say about how to interpret those results.

While it may not be initially obvious, results in figure 4.5 [showing the relationship between steak demand and food values] can be interpreted as providing evidence about people’s beliefs about (or perceptions of) steak. Suppose an individual highly values taste. Figure 4.5 shows that such an individual will tend to choose more steak. As a result, it must be that steak is perceived to be highly tasty. By this line of reasoning, figure 4.5 suggests that consumers, on average, perceive steak to be convenient, tasty, attractive, and novel but they also perceive steak to be poor for animal welfare, nutrition, and environment while also being expensive.

There's a lot more in the report.

Disruptive Trends in Food and Agriculture

In the past couple weeks, I've had several opportunities to engage with some forward looking farmers and agribusiness executives, and a common theme seems to have emerged around many of the conversations: what are the issues or food and agricultural technologies on the horizon that could be potentially disruptive for the current incumbents?  

1) Block chain technology.  This isn't bitcoin, but rather the underlying technology that facilitates bitcoin trades, which could be applied to many other industries.  This Reuters article from earlier in the week, for example, indicates, "A cargo of U.S. soybeans shipped to China has become the first fully-fledged agricultural trade conducted using blockchain."  The thought is that blockchain technology might prove to be a mechanism that can more rapidly disseminate many types of information about trades (the Reuters article mentions the "sales contract, letter of credit and certificates") more widely and rapidly.  Big players like Walmart and IBM are also talking about using blockchain to improve traceability and food safety.

2) Plant-based and cellular-based protein.  This is a topic I've written about many times in the past (e.g., here or here).  What's changed is the high level of investment flowing into this space, including by companies like Tyson and Cargill.  Moreover, there are now products from companies like Impossible Foods, Beyond Meat, JUST, and others that are actually in the market.  If sales ramp up, what are the impacts on producers of current animal feeds (primarily corn and soy)?  What are the new agricultural inputs for these plant-based meat/egg/dairy alternatives? 

3) CRISPR.  Again, the basic science isn't necessarily new,  but there are new applications coming on board (non-browning apples, hornless Holsteins, etc.) and potential changes in the regulatory landscape that could accelerate (or decelerate) adoption and consumer acceptance.

4) Agricultural analytics.  This includes precision agriculture, sensing, big data, drones, modeling, etc.  Yes, these have been around for a while and there have been many discussions about data ownership and rights, but there is a sense that the data and technology have moved to a point where some adopters may be able to start gaining a competitive advantage. 

5) Online food buying.  Will Amazon do to the food supply chain what they've done in other industries?  Walmart is also making big moves into this space.  What are the implications for traceability, tracking, and vertical market coordination?

6) Trade.  Agricultural trade has a big impact on US agriculture, and it appears there may be changes in trade policy on the horizon. 

What have I missed?

Don't Want to Eat Pink Slime? Would You Even Know?

It's hard to believe it's been almost five years since the finely textured beef (aka "pink slime")  scandal broke.  To briefly re-cap, by 2012 it had become an industry standard to include finely textured beef with other beef trimmings to make ground beef.  The process enabled food processors to add value, cut down on waste, and increased the leanness of ground beef in an affordable manner.  But, a series of news stories broke, which caused public backlash against the process, and ultimately led to the closure of several plants that produced finely textured beef.  In 2013, I wrote about my visit to BPI, one of the largest producers of lean finely textured beef (this summer, ABC settled a multi-million dollar lawsuit brought by BPI regarding ABC's coverage of the issue).  I devoted a whole chapter of my 2016 book, Unnaturally Delicious, to the issue.  I'll also note, for some aspiring journalist out there,  that I can imagine a highly compelling a book-length treatment of the saga.

Back to the heart of the story, must of the public backlash presumably came about because the public was worried about taste or safety of ground beef made with finely textured beef.  In the monthly Food Demand Survey (FooDS), we've been running for almost five years, we ask about perceptions of the safety of "pink slime" and of "lean finely textured beef".  The data suggests neither are top safety concerns.  The most common answer is that people are "neither concerned nor unconcerned" about the safety of these issues (for lean finely textured beef, the average response is actually in the direction of "somewhat unconcerned").

Well, what about taste?  People may think "pink slime" tastes bad, but what would happen in a blind taste test?  Along with several of my former econ and meat science colleagues at Oklahoma State University (Molly Depue, Morgan Neilson, Gretchen Mafi, Bailey Norwood, Ranjith Ramanathan, and Deb VanOverbek), we conducted a study to find out.  The results were just published in PLoS ONE.  Here's what we found.

Over 200 untrained subjects participated in a sensory analysis in which they tasted one ground beef sample with no finely textured beef, another with 15% finely textured beef (by weight), and another with more than 15%. Beef with 15% finely textured beef has an improved juiciness (p < 0.01) and tenderness (p < 0.01) quality. However, subjects rate the flavor-liking and overall likeability the same regardless of the finely textured beef content. Moreover, when the three beef types are consumed as part of a slider (small hamburger), subjects are indifferent to the level of finely textured beef.

So, a burger made with 15% finely textured beef is as tasty or tastier than a burger without finely textured beef.  If people knew this, would it have changed their reaction to the Jamie Oliver show or the 2012 ABC News stories?   

How Innovative are Food Sellers?

I tend to think of the food and grocery business as hyper-competitive with many new product introductions and failures.  For example, data from the USDA ERS suggests roughly 20,000 new food and beverage products come on the market each year.  That seems like a lot.  But, compared to what?

I ran across a presentation that Austan Goolsbee and Pete Klenow recently gave at the ASSA conference in Philadelphia.  They use online price and quantity data collected by Adobe Analytics.  Their main objective was to construct price indices to compare against the official government consumer price index measures.  Those results are interesting, but I was intrigued by some of their other findings regarding new product introductions and exits.

Below are two slides they created that challenged my prior beliefs about the food and beverage category.  

If I'm understanding this correctly (and I may not be), I think the data below suggests that 69.8% of the total sales in the apparel category come from newly entering products.  Moreover, 29.5% of total sales in the the apparel category come from products that are soon to exist the market.  Other categories with high levels of "churn" and new product introductions are "other goods and services", ICT (which I believe is information and communications technology), and recreational goods.  For food and beverages, "only" 19.5% of sales are by new entrants, and 8.5% of sales are by soon-to-be-gone products. By this measure, the food sector seems less dynamic than others. 


Their next slide conveys the same information in a different way - by measuring the percent growth in sales-weighted variety.  By this measure, variety in food and beverages has only grown by an average of 1.2% per year, and the only category with a smaller growth in variety is medicines and medical supplies. By contrast, there is an average 18.3% growth in vareity in the aparel cateogory.  


One downside to these data is that they only reflect online sales.  Perhaps there is more "churn" and new product introduction in brick-and-mortar grocery stores than online?  Either way, this is interesting food for thought.