Are small, local supply chains more resilient? The 2020 meat plants

By late April 2020, more than 30 meat-processing plants in the United States had closed because of outbreaks (Hobbs, 2021). Wholesale meat prices rose while livestock prices at the farm fell (Hobbs, 2021). Many people asked at the time: if chains were smaller and closer to the market, would they have coped better? This post reads Hobbs (2021) to see what the evidence says, then turns the question into a worksheet for a small business. The framework used is Efficiency versus resilience lens; the case summary is on the library page.

What happened to the meat plants in 2020?

Processing capacity dropped sharply and quickly. In the United States, daily beef and pork processing ran about 40% below 2019 in the last week of April and the first week of May 2020, and federally inspected slaughter averaged about 22% below 2019 from early April to early June (Hobbs, 2021). On health, the CDC reported 9% of meat-plant workers infected from late March to May 2020, with outbreaks at 239 meat and poultry facilities (Hobbs, 2021).

In Canada, three plants hold 85% of beef slaughter capacity, and capacity fell by more than 60% in late April and early May (Hobbs, 2021). Livestock prices fell and wholesale prices rose; retail beef returned to near its pre-Covid level by June 2020, and the hog market settled in late summer or early autumn (Hobbs, 2021). These are North American figures; I have no equivalent for Vietnam's meat chains.

Did the food chain collapse?

Not in Hobbs's assessment. He writes that food chains "performed remarkably well" against the pandemic's shocks, with the main vulnerabilities being worker health in meat plants and in fruit and vegetable production (Hobbs, 2021). One thing that helped was governments exempting people who move and distribute food from border restrictions (Hobbs, 2021).

Do small, local chains do better?

There is little evidence that they do, and Hobbs gives three reasons (the grouping into three is mine). First, a simulation by Ma and Lusk (2021), in his account: a concentrated sector keeps output high better, a dispersed one keeps output above a minimum better, the differences are small, and a dispersed sector lowers welfare through lost economies of scale (Hobbs, 2021). Second, small plants face the same labour vulnerabilities and may be more labour-intensive (Hobbs, 2021).

Third, according to Thilmany and colleagues (2021), whom Hobbs cites, local and regional actors are more nimble but lack scale and may be vulnerable to local disruptions such as weather (Hobbs, 2021). Hobbs's conclusion, written in March 2021: caution should be exercised in calling for dispersed, local chains, since there is little evidence a dispersed system would have done better (Hobbs, 2021). He also cites Chenarides and colleagues (2021) on investing in flexibility in normal times, which can be worth something as a real option (Hobbs, 2021); that is the direction I take in the worksheet below.

What does "resilience" mean, and why separate its kinds?

Hobbs uses two definitions. Ponomarov and Holcomb (2009), as he quotes them, see supply chain resilience as the adaptive capability to prepare for unexpected events, respond to disruptions and recover by maintaining operations at the desired level (Hobbs, 2021). Tendall and colleagues (2015) define food system resilience as the capacity over time, at multiple levels, to provide sufficient, appropriate and accessible food to all in the face of various and even unforeseen disturbances (Hobbs, 2021). I read only the abstracts of the two original papers (Ponomarov and Holcomb, 2009; Tendall et al., 2015).

Hobbs also separates inherent resilience, a chain's ability in normal times, from adaptive resilience, its ability in abnormal times (Hobbs, 2021); see inherent versus adaptive resilience. For an owner the companion question is: are you relying on what the chain has in place in normal times, or on your ability to adjust when something breaks?

Reading the two definitions together is my own step. The first is about a firm keeping operations at a desired level through a disruption; the second is about whether food reaches people. For a small business the first is the usable yardstick, provided you set the "desired level" before the incident, for example "at least 60% of daily meals within three days of a closure", and then test whether the standby option can actually deliver that level. Without a stated level, "resilient" is only a mood. Hobbs's own examples are on a national scale, so what a single firm should aim for has to be your decision, based on your contracts and your customers' tolerance.

Which worksheet helps decide on a standby option?

I propose five questions for each single point of failure, a point whose loss stops the whole chain; this worksheet is mine, not Hobbs's. One: which point stops the whole chain? Two: how much efficiency does it give in normal times? Three: if it fails, how many days and how much money do you lose? Four: what does a standby option cost each year? Five: are you relying on inherent or adaptive resilience?

A hypothetical example, every figure invented by me. "Bếp Ví Dụ" is a central kitchen cooking 4,000 meals a day, 300 days a year, for offices. Each meal contributes 8,000 dong of profit after variable costs. The kitchen is a single point of failure. Suppose the chance that it must close for 14 days in a given year is 10%.

The 2.4 billion dong is an invented assumption: the profit from one year of sales to the 25% of customers who leave for good after the closure (25% × 9.6 billion dong of annual profit). Counting only the 14 days of lost profit, a standby kitchen is not worth buying: an expected loss of 44.8 million is far below 300 million. Add the assumption that 25% of customers leave for good and the two sides are nearly equal (285 million against 300 million). The decision then rests on two numbers you cannot know for sure: the probability of closure and the share of customers lost. The sum also ignores tail risk: a closure much longer than 14 days could end the business, and an average expectation does not show that. Before buying, compare a cheaper option, such as an agreement to rent a shared kitchen by the day when needed.

The point shared with Hobbs: the answer does not lie in the slogans "split it up" or "go local", but in what your particular single point of failure costs when it fails against what protection costs each year. Two related posts for owners: efficiency or responsiveness and two ways a supply chain breaks in Britain.

Going deeper: what I read and what I did not

I read Hobbs (2021) in full, in the public copy on PubMed Central; it was written in March 2021, so it has no later data. Every figure on the meat plants, the studies by Ma and Lusk, Thilmany, and Chenarides, and the two definitions of resilience come from Hobbs's account of them. I did not read those original studies. I read only the abstracts of Ponomarov and Holcomb (2009) and Tendall and colleagues (2015), so I cannot judge the two definitions in their original context. My copy has no page numbers, so I cite the article as a whole.

The five-question worksheet and the kitchen "Bếp Ví Dụ" are mine, with invented numbers: 4,000 meals, 8,000 dong, a 10% probability, 25 million a month, 25% of customers lost. I have not tested them on a real business, and there is no Vietnamese data in this post.