Supply Chain Resilience

Sustained, strong demand is, I perceive, the best guarantor of resilient supply chains. High volume, high velocity contemporary supply chains demonstrate rather amazing creativity and adaptability when pull signals keep singing. (The reverse can also be true, when pull is muted, push is demoralized and distracted.)

As such the monthly retail sales and personal consumption expenditure reports are among our best indicators of overall supply chain fitness. Below is a helpful chart condensing this morning’s retail sales report for June.

The Census Bureau delivers these details, “Advance estimates of U.S. retail and food services sales for June 2024, adjusted for seasonal variation and holiday and trading-day differences, but not for price changes, were $704.3 billion, virtually unchanged (±0.5 percent) from the previous month, but up 2.3 percent (±0.5 percent) above June 2023. Total sales for the April 2024 through June 2024 period were up 2.5 percent (±0.5 percent) from the same period a year ago. The April 2024 to May 2024 percent change was revised from up 0.1 percent (±0.4 percent)* to up 0.3 percent (±0.2 percent). ” (More and more.)

Category-specific outcomes were mostly better than the overall. Used car sales took a huge hit in June because of the cyberattack on back office systems serving that sector. Excluding disrupted auto sales and lower priced gasoline, total retail sales increased 8 percent month-over-month. Despite the surprisingly strong increase, both wholesale and retail inventories increased slightly. This is not a stress test, this a healthy habit of day-after-day exercise.

The second chart below — showing sales at grocery stores and eating-out places — suggests to me a consumer population that since last November has found a balance of careful shopping and occasional fun that reasonably reflects recent wage trajectories.

Contemporary supply chain management works to minimize costs of excess capacity and inventory. The goal is to accurately anticipate demand so that supply is pushed Just-In-Time to precisely where consumer pull is present.

Because it is less costly to make and buy in bulk and precise prediction is not always possible, there is usually some upstream buffer stock. Mid-stream capacity is purposefully tight. Trucks and drivers are expensive, what you’ve got you want to fully utilize. Each firm’s last mile freight capacity is shaped to deliver to highly probable demand at full-flow velocity. Downstream stocks are intended to be just enough to supply each customer’s demand between deliveries.

Dollar General is currently — perhaps perpetually — in the process of supply chain streamlining. During a May earnings call CEO Todd Vasos said, “… our supply chain teams are primarily focused on ensuring deliveries are on time and in full and we are reducing the amount of inventory we carry… Our top priority in this area continues to be improving our rates of on-time and in-full truck deliveries, which we refer to as OTIF. Our distribution and transportation teams have taken aggressive action to improve their service to our stores, and these efforts have led to significantly higher OTIF levels compared to the same time last year.” In the same call the CFO emphasized, “Notably, total nonconsumable inventory decreased 19.1% compared to last year and decreased 22.5% on a per store basis. The team continues to do great work reducing our overall inventory…” (More and more.)

Depending on products, customer preferences, and investment resources (including human judgment), other firms will emphasize slightly different priorities, but delivering optimal flows at minimum cost is a nearly ubiquitous objective. The result is less waste, reduced costs, and more sustainable margins. This structural efficiency can, however, reduce options when demand suddenly shifts. The more volatile consumer pull, the less effective related push (for example 2020-2021).

In advance of a hurricane or blizzard or such, demand will surge. Grocers and fuel operators in particular have experience and protocols to increase supply into the face of anticipated demand. More deliveries are scheduled and, for grocery, the product mix can dramatically shift. The goal is to serve both pre-event and post-event needs. There is, however, seldom time or opportunity to engage many more trucks or drivers. The focus has to be on moving more with what is already on hand. To achieve this waivers of freight regulations can be very helpful, especially in advance. When a hurricane or blizzard is forecast, the supply chain “disaster” usually starts about 72 hours before the actual winds pick up. Allowing trucks to carry slightly more than usual and truckers to stay-on-the job slightly longer than usual will improve velocity and, therefore, volumes available.

Texas authorized Hours-Of-Service and weight waivers almost 48 hours before Hurricane Beryl’s Monday landfall.  But for some reason these waivers were NOT widely communicated. Suppliers sent me notes describing the public sector communications process as “terrible”, “convoluted”, and “pathetic”. The frustrated reaction was, in part, anchored in disappointment. It is much easier to get intra-state waivers in a hurricane-ready state than federal-plus-multiple-state waivers from less experienced players. Freight carriers expect the feds and various southbound states to be passive and/or confused before a Florida landfall. Texas is very experienced with hurricanes. Texas is big enough that their single set of intra-state waivers will usually maximize flows across the entire Texas Triangle (Austin, Dallas, Houston, in-between, and near-by). Texas usually deploys and communicates much better than average. But not in advance of Beryl.

I’m sure this will be part of the Texas after-action for Beryl. Even more important in my judgment: If Beryl had been forecast at CAT3 or above, could freight waivers have been released — and effectively communicated — before Saturday, even as early as Thursday morning? What about when similar waivers are needed for North Carolina, South Carolina, Georgia, Alabama, Mississippi, Louisiana, and Tennessee to increase volumes and velocities when a major hurricane is forecast for south Florida? Or a major Northeast Blizzard? Or other rapidly emerging threats?

Supply Chain Resilience actions can often be characterized by one of four risk-oriented choices: Transfer, Avoid, Reduce, and/or Accept (TARA). Risk abundantly unfolds. We will experience bruising or worse. But with forethought and action, most risks can be mitigated.

For example, potential loss-of-life from storm surge can be avoided by evacuating surge-prone places. Potential property damage from hurricanes can be reduced by purposeful location, architectural, engineering, and construction choices. Property insurance can sometimes (in some places) be purchased to transfer the risk of financial loss to the insurance carrier and their underwriters. We often make decisions — explicit and implicit, commercial and personal — to accept some level of risk. With enough free cash-flow some can self-insure, which is basically a more-thoughtful-than-usual approach to risk acceptance.

Risk transfer is increasingly expensive — and treacherous . Over the last decade underwriting losses have accumulated. According to A.M. Best, “In 2023, the [US] industry recorded a net underwriting loss of $21.6 billion, following a $25.8 billion loss in the prior year.” The marginal improvement was due to 2023’s comparatively light hurricane season. Overall global losses — not just insured losses — have continued to exceed historical averages (see chart below).

Insured losses are especially acute in homeowners and vehicular insurance lines (more), but the commercial property insurance market is also morphing as disasters increase in frequency, virulence, and cost.

According to Travelers Insurance, “In the last four years, these [catastrophic] events have caused annual insured losses of more than $100 billion globally. In 2023, total insured losses globally were an overwhelming $118 billion. Severe convective storms (SCS) represented 58% of the losses globally, and in the U.S., six of the 10 most expensive events were SCS events.”

With more destruction and higher claims, the cost of insurance has increased. This increased expense has — as surging demand often does — tended to result in system-shedding of the most vulnerable or peripheral potential consumers. The Consumer Federation of America reports, “One in thirteen homeowners across the United States are uninsured (7.4 percent), equivalent to 6.1 million homeowners. Homeowners making under $50,000 a year are twice as likely as the general population to be uninsured (15 percent).”

Inflation is a contributing factor for increased losses (and another common outcome of demand surges). A 2022 Harris Poll found that, “only 43% of U.S. business owners have increased their business insurance policy limits to account for inflation. These updates were influenced by things such as increases in the costs of goods and services, higher wages, and increased interest rates for loans, all of which may impact the cost to rebuild or resume operations after a claim.”

Reinsurance News reports, “Over the past five years, commercial insurance lines have faced challenges due to adverse prior year loss reserves, latent liabilities, and increased competition in longer-term coverage areas… Increasing reinsurance costs, particularly for catastrophe-exposed lines, along with tightened capacity in affected property markets, have necessitated higher account pricing that contributed to higher calendar-year premium totals.” In an admittedly self-interested comment published by Risk&Insurance magazine, a truth was nonetheless articulated, “Driven by natural catastrophe and extreme weather, a surge in inflation, inaccurate property valuations and more, the commercial property insurance market has seen several consecutive years of increasing premiums. The world is a risky place. Protecting assets and reducing risk in today’s volatile and uncertain environment can feel like a daunting task for any business.”

As risks increase and transfer costs follow, network resilience is stressed by relative decreases in overall risk management investments which can be hidden until an amplified shock hits the under-invested place and people waaay too hard.

+++

July 11 Update: Hurricane Beryl did extensive damage across the Caribbean (more and more). Reuters is reporting a wide range of damage estimates related to the hurricane’s hit on Texas. Reinsurance News has highlighted the Karen Clark & Company estimate of $2.7 billion for Texas-specific insured losses. [Cost-related — but not Beryl related — residential insurance analysis from the NYT.]

July 20 Update: According to the Houston Chronicle, “The Texas Windstorm Insurance Association (TWIA) said in a meeting this week that it anticipates draining half of its catastrophe reserve funds to cover windstorm damage payouts. The TWIA is an insurer of last resort for many Texans, giving policies only for wind and hail damage to customers who have been denied coverage in the public market. Chief Actuary Jim Murphy said that the agency expects to use about half of its over $450 million catastrophe fund to pay for the 16,000 claims it has already received. That figure could go up to 20,000 claims for an estimated more than $200 million in payouts. ”

[July 1 Update below] Some readers found my June 10 post on hurricane season, “a bit too thin” or “… “too fast”… or “too big-picture” (among other comments, including some positive). So, here’s a try at something meatier, slower, and more tightly focused. The original post was organized around threats, vulnerabilities, and consequences. Looking at Hurricane Ian’s consequences I argued, “Because midstream capacity concentrations were not hit, flow continued downstream to concentrations of demand. The harder the hit upstream the more disruption (or worse) downstream.”

The threat of a major hurricane — Category 3 and above (more than 111 miles per hour) — in the United States is concentrated in Florida, Texas, and Louisiana. Since 1851 three-quarters of major hurricanes recorded on the mainland United States have involved these three states. (Puerto Rico and Guam are much smaller targets but experience even more frequent cyclones per square mile.) The first map below only shows Category 4 and 5 hurricanes since 1924.

In most of the United States supply chains will usually be resilient (not invulnerable, but adaptable) unless key capacity concentrations are lost or throughput is significantly reduced for an extended time. Two well-known examples are when the Colonial Pipeline was disrupted by a cyberattack and when contamination halted production at Abbott’s Sturgis, Michigan infant formula facility. In each case a “Pareto Proportion” of supply for specific sources of demand suddenly stopped.

Similar concentration risks exist in a wide array of networks. Food tends to be most heavily concentrated at the distribution level.(Food production concentration can also be intense, as the infant formula example demonstrates. But in most food categories there are abundant replacement products.) Fuel is highly concentrated at both production and distribution levels. The second map below shows refineries located in “hurricane alley”. The third map shows refined product pipelines. (In Florida fuel flows mostly depend on maritime deliveries to Port Tampa Bay — more than 40 percent — and Port Everglades — about one-third of state fuel consumption.) Local fuel distribution capacity is mostly fixed by the number of nearby racks, regionally anchored fuel tanker trucks, and their drivers. Freight flows — especially long-distance flows — are concentrated in a comparatively few channels. For example, the fourth map below shows how freight flows depend on a few Interstate highway corridors. I-10, I-75, I-35 and I-95 would be especially tough to replace.

Over 70 percent of food consumed in the United States is carried by truck. In most of this hurricane prone region some three-party combination of Walmart, Publix, HEB, Kroger, C&S Wholesale, Albertsons, and Associated Wholesale Grocers distribute a Pareto Proportion of food consumed. Their Distribution Centers are often clustered (see the last map for an example). Their inbound and outbound trucks usually take very similar routes. According to the Commodity Flow Survey, Freight Analysis Framework, and other sources roughly one-third of food products in this region originate in the host state and another 10 to 15 percent originate in a neighboring state. Roughly half of food consumed comes from farther away. (The FEWSION project is a good source for aggregated food flow data and helpful measurement tools.)

Right now, the electrical grid is humming. Fuel pipelines and racks are flowing. Interstates are open. Truck stops are operating. In late June, gasoline and diesel inventories for this region are in good shape. According to FreightWaves, trucking capacity in the Southeast US is tighter now than in recent years — but perhaps better matched with consistent demand. According to DAT, the demand for refrigerated vans — especially important for food flows — is tighter in the Southeast than in most of the rest the United States. But this is “tight” as in tailored, not a death-grip. Grocery distribution centers are receiving inbound from near and far while discharging daily into their retail networks.

Good news: These food, fuel, and freight concentrations have enormous capacity. When and where this capacity persists it is fundamental to serving the needs of those hit hard by hurricanes or other disasters. Bad news: if these capacity concentrations are seriously disrupted alternatives can be few and far between…

The hurricane forecast for the Atlantic is foreboding. Ocean temperatures are hot (here and here). As of Thursday morning, June 27, the National Hurricane Center says, “A tropical wave located several hundred miles west-southwest of the Cabo Verde Islands continues to produce disorganized shower and thunderstorm activity. Environmental conditions are forecast to be conducive, and development of this system is anticipated. A tropical depression or tropical storm is likely to form this weekend several hundred miles east of the Windward Islands while the system moves westward at 15 to 20 mph.”

Ready or not.

+++

July 1 Update: Since Thursday that “Tropical Wave” has become Hurricane Beryl. The Weather Channel reports, “B​eryl is the easternmost hurricane and first “major hurricane” (Category 3+) to form in the tropical Atlantic during the month of June.” In the next 24 hours Barbados, St. Lucia, St. Vincent and the Grenadine Islands, Grenada and Tobago will be hit hard. Below is how the models set-up potential action for late this week. Depending on Beryl’s track, very different capacity concentrations are potentially vulnerable. More on July 2 and the Washington Post reports, “Beryl strengthened into Category 4 a week earlier than any storm of that strength ever observed, breaking a record set by Hurricane Dennis in the hyperactive 2005 storm season. It also became the fastest-strengthening storm on record before the month of September. This kind of early-season activity in the area is a strong predictor of a large tally of tropical storms by late fall…”

According to the USDA, during 2024 the world is leaning toward a bit less wheat, corn, and soybeans. But existing stocks will mostly fill the gap. There will likely be a bit more rice. But whether up or down the shifts are under one percent on huge volumes. Large-scale, short-term weather and crop conditions look okay in most high-production places (see map below). According to the World Meteorological Organization, a transition from an El Niño to a La Niña weather pattern is underway. In the next several weeks WMO forecasts a “large-scale cooling of the ocean surface temperatures in the central and eastern equatorial Pacific Ocean, coupled with changes in the tropical atmospheric circulation, namely winds, pressure and rainfall. The effects of each La Niña event vary depending on the intensity, duration, time of year when it develops, and the interaction with other modes of climate variability. “

Last week the US Energy Information Administration wrote, “We forecast U.S. crude oil production will grow by 2% in 2024 and average 13.2 million barrels per day (b/d) for the year and a further 4% in 2025. If our forecast is realized, U.S. crude oil production would set new annual records in both 2024 and 2025.” Global oil futures — bouncing around the low to mid 80s — imply widespread expectations for sufficient supply (see Brent Crude chart below). A provocative new long-range forecast by the International Energy Agency offers:

Based on today’s policies and market trends, strong demand from fast-growing economies in Asia, as well as from the aviation and petrochemicals sectors, is set to drive oil use higher in the coming years, the report finds. But those gains will increasingly be offset by factors such as rising electric car sales, fuel efficiency improvements in conventional vehicles, declining use of oil for electricity generation in the Middle East, and structural economic shifts. As a result, the report forecasts that global oil demand, which including biofuels averaged just over 102 million barrels per day in 2023, will level off near 106 million barrels per day towards the end of this decade. In parallel, a surge in global oil production capacity, led by the United States and other producers in the Americas, is expected to outstrip demand growth between now and 2030. Total supply capacity is forecast to rise to nearly 114 million barrels a day by 2030 – a staggering 8 million barrels per day above projected global demand... (More and more.)

The IEA forecast depends on global grid capacity expanding as quickly as needed between now and 2030. This is possible, but will not happen without serious challenges — and occasional failures — along the way. There are both near-term and long-term gaps to fill, especially where demand is growing fast. ERCOT, the Texas grid operator, is an especially colorful canary in this coal mine. Spectrum news recently reported, “ERCOT says there is a 12% chance of rolling blackouts this August. Officials are also warning that demand will increase exponentially over the next several years… A 40-gigawatt increase from one year to the next in the five-year horizon. That’s effectively almost doubling the peak demand of the ERCOT grid in about six years.” (More and more and more. )

Fuel costs associated with crude at under $90 per barrel tend to result in less freight cost volatility (and related structural stresses). The gradual normalization of Panama Canal transits incrementally smooths global flows. Suez canal transits are, of course, still suppressed and down nearly two-thirds from last year. Still, despite various frictions (here and here and here), overall global trade volumes are up over slow 2023 flows (here and here). Given otherwise strong economic data, there is no particular reason to expect the real-value of US goods imports to fall far in the still underway second-quarter. The Global Supply Chain Pressure Index continues to show lower than average measures. The Cass Shipment Index has been softening but remains in its mid-range. DAT Trendlines are mixed, but big early summer increases in year-over-year load-to-truck rates are meaningful.

Millions of people are largely excluded from these robust flows. This is not, however, caused by upstream capacity constraints. Places on the periphery of big midstream channels are more likely to have problems. But absence or paucity of effectual downstream pull is the most common culprit. There are significant geo-political, climate-related, and other natural risks that could catastrophically subvert upstream capacity and midstream velocity with dangerous downstream results. But as the second half of June leans into the second half of 2024, global flows are strong.

+++

June 24 Update: Using the same or similar information sources, Peter Goodman at the New York Times offers a very different interpretation than above — or that offered in my May 27 overview — regarding freight flows. We both perceive increased friction, related congestion, higher prices, and plenty of prospects for additional risks. I tend to see high capacity complex networks morphing with volatile context. Mr. Goodman seems to see a highly concentrated, engineered system being manipulated in unsustainable ways. There is certainly some of each (and much more) going on. What is predominant? What is the straw that might break the camel’s back? What flutter of butterfly wings will facilitate persistent flows?

June 25 Update: I wonder if some Bloomberg editor explicitly decided to challenge Peter Goodman’s argument. In any case, this morning there is an alternative interpretation of supply chain stress available here. The Bloomberg piece channels Jason Miller explaining that upstream capacity is well-calibrated with (even a bit excessive considering) current demand… and midstream channels have the ability to maintain necessary flow velocity. This is suspiciously close to my own perception, so I may be hearing what I want to hear. Worth reading both points-of-view.