*CW: meat consumption, animal death.
We are a few weeks away from Thanksgiving, and the number one rule of supply chain Substack is to write about turkeys (I don’t make the rules), and as you might expect, this “supply chain” is already having issues.
Long-time readers of this newsletter know that the only constant in supply chains is that there is no constant.
And indeed:
“Prices for whole birds (16 pounds) are 28% higher than one year ago, with prices being up 0.7% in the last week. In September, turkey prices accelerated by 1.3%, according to the US Consumer Price Index – with consumers paying for all kinds of “uncooked turkey” 17% more than one year ago.”
And the reason? While inflation plays its role, it’s really the avian flu outbreak:
“Federal officials say the highly pathogenic avian flu has killed 47.72 million commercial poultry in the U.S. this year, 7.4 million being turkeys. One of the latest outbreaks was at a commercial turkey farm in Pennsylvania that killed 15,500 turkeys.”
But if you’re wondering whether you'll actually have a turkey for Thanksgiving, a shortage is not anticipated:
“‘The shortage was worse throughout the summer than it is now, but the prices are still high. We don’t anticipate there being another shortage, but the size of the birds may be smaller. People may have to make the choice to get a smaller bird,’ said Dan Piron, a manager at Green Hills Farms Grocery in Syracuse.”
This is an interesting type of scarcity that we rarely see. For instance, when dealing with the great kettlebell shortage, there wasn’t an option for a smaller kettlebell, or for toilet paper with fewer layers, or coffee with less caffeine … you get the point. Usually, when faced with a shortage, the product in question is or isn’t available.
As usual, the task of this newsletter is to delve deeper and understand the operations behind this news. This year I’m also making sure my readers can make wittier (or “geekier”) comments at their Thanksgiving dinners. And for my international readers who don’t celebrate Thanksgiving, I hope you find this interesting regardless.
The Turkey Vendor Model
In most of the previous supply chain models we’ve studied, there is underlying demand uncertainty, which makes supply chain decisions risky.
In the turkey supply chain, demand uncertainty is not that much, as we can see from the following graph:
So the primary reason that the turkey supply chain is different is that there’s a problem of uncertain yield, rather than uncertain demand.
To some extent, the two main questions this supply chain faces are that of planning at the beginning of the season, and that of mitigation (or reaction) once a “shock” has been observed, in the middle of the season.
What do I mean by this? Knowing that there is always (and potentially now more than ever) a risk of disruption affecting the yield (how many turkeys will survive), the question is how many turkeys should be raised? And when a health shock is observed, the question is what measures can be taken to minimize the downside (or maximize the upside) for the farmer?
Let’s start with the first issue: the uncertain yield. Supply chains with uncertain yields are not new, but they’ve been studied much less than those with demand uncertainty. Some examples of supply chains with uncertain yield are the development of flu vaccines, and the manufacturing of semiconductor chips.
We all know the canonical newsvendor model (which I’ve used before), where demand is uncertain, and the producer must decide how many products to produce knowing there might be an overage or underage. We all also know the critical fractile solution —the optimal service level should be the fraction of the underage cost to the sum of the under and over costs.
But newsvendor problems with uncertain yield are much harder.
In fact, their solution is quite complex.
An interesting paper studies not only the problem, but also how good humans are at solving it, and …
“...we present experimental results that show how difficult newsvendor decisions under supply uncertainty are for human subjects. In our experiment, the control group replicated a well-known newsvendor experiment whereas the test group faced additional supply yield uncertainty. Comparison of these results shows that under low-profit condition, subjects are able to incorporate supply uncertainty quite well in their decisions. Under high-profit condition, the deviation from the optimum is much more significant.”
But I would say that turkey growers are doing a good job incorporating this yield uncertainty given that the “average net profit margin of 2.9%, turkey producers, which would be grouped in the ‘Meat Products’ sector of U.S. industries, would rank 171 out of some 215 industrial sectors.”
The Mitigation Dilemma
But fast forward to the situation we are in now. If you are a grower that has not yet been affected by the outbreak, what should you do?
The problem is the following:
“The current strain of avian flu is more contagious and causes severe illness, according to Watt Global Media, which follows the poultry industry. When one bird in a flock is found to have avian flu, the entire flock is contaminated and cannot be kept alive.”
So as a farmer, you have an interesting dilemma:
You can either process and freeze the turkeys you have now, making sure you have enough turkeys to sell in time for Thanksgiving, but: the price of a frozen turkey is lower than that of a fresh one, and the size of the turkey will be smaller (since it didn’t have time to grow). Nevertheless, the outcome (yield) is somewhat certain.
And indeed:
“Many commercial turkey producers butchered and froze their flock earlier in the year to avoid losing them to the avian flu, Piron said, so the birds are not as big as consumers may expect.”
Or, you can choose to bet that none of your birds will get the flu, the upside being a higher profit, both in terms of the size and price of the bird (fresh vs. frozen). However in this case, there is risk involved.
The trend of bigger (and fewer) turkeys has been brewing for many years.
The problem the farmers face can then be described as a decision tree.
Process and Freeze vs. Keep Growing:
Let’s assume that you have 15,000 turkeys.
In the process and freeze scenario, you should expect a whole frozen turkey, weighing between 8 to 16lbs, to set you back roughly $1.99 per pound.
If you keep growing, fresh turkey breast meat costs an average of $6.50 per pound, and we assume that your full-grown turkey weighs around 30lbs. Just for comparison, according to the New York Post, in 2020 turkey breast cost just $2.00 per pound.
So, the two scenarios can be translated into:
15,000 x 1.99 x 12 (average weight) vs. 15,000 x 6.5 x 30 x likelihood of getting the avian flu on the farm.
Quick math will reveal that if the likelihood is below 12%, you should continue growing them. If the likelihood is above, you should process the birds right away.
What determines the likelihood of your birds getting the avian flu?
The size of the farm. Of course, getting the flu is highly correlated since the birds all feed, breathe, and sleep together, but still. Let’s assume that the likelihood of every turkey getting avian flu is somehow independent of its farm-mates.
Using a very simple calculator for Bernoulli trials (assuming all turkeys are independent), even if the probability of each turkey getting avian flu is around 0.0009%, the probability that at least one turkey in the farm gets avian flu is 12%. Reminder: if at least one turkey gets the flu, you have to kill the entire flock.
(Note: Check my math. As I tell my students, I make many mistakes. Some are intentional to keep you awake, and some are not).
To me, this seems like a very low probability threshold and a really tough dilemma for the farmer.
And indeed:
“Farmers are working very hard and we are liaising with the Government to explore ways to support producers like being able to slaughter earlier, freeze and then sell defrosted. Those who buy a free range turkey may see fewer of those specific birds available due to the impacts of bird flu.”
This issue is not unique to the US. The UK has suffered equally, and extra measures are being taken to help those who want to keep their turkeys alive and healthy. There are now mandatory measures requiring turkey quarantine:
“The UK government explains that ‘evidence shows’ that housing birds reduces the risk of the birds contracting the flu.”
“First, they came for the people, and I did not speak out” … Well,… Nevermind.
Just know that if you opt for a non-frozen bird, you shouldn’t expect it to be very social.
In case you’re wondering, I love Thanksgiving (a newly adopted holiday since we don’t celebrate it in Israel). I’m not a huge turkey fan, but I do follow tradition. However, my favorite Thanksgiving dish is sweet potato casserole, so if the tradition ever changes and eating turkey is not a Thanksgiving thing anymore, I won’t be the one to protest.
P.S. The art is created using the generative AI site lexica.art, using the phrase “Thanksgiving Turkey Supply Chain.” Pretty good. You don’t want to see the other results though.
Curious to imagine when Beyond or Impossible products can significantly start to appear on the Thanksgiving table as a way to offset the smaller birds!
Additional strategies for mitigating some of the risks in this fowl situation:
1. Bi-modal: process + freeze X% of the flock AND grow to full size (1-x)% of the flock - locks in a base level of gains while still getting exposure to gobble up the upside. By reducing the flock size, you also decrease the density and therefore risk of avian flu.
2. Insurance. Just as with crop-insurance, I would imagine the turkey producers could insure against avian flu outbreak
3. Hipster Tom: Sell members of your flock to individual families during the summer (at a premium) and then they raise the bird themselves.
4. Separation: split a single farm into separate farms so you don't lose all of your birds if one group gets infected.