Supply Chain Resilience

Inflation-adjusted July Personal Consumption Expenditures on Food-At-Home once again increased slightly (see chart below). If we continue to see these “slight” increases for a couple more consecutive months I will have to retire my use of “flat” to characterize the current slope. But this July our real PCE for food was $1032.4 billion while last July our real PCE was $1027.9 billion. That’s still flat enough for me. What’s less than $5 billion among friends and fellow citizens?

Here’s how Bloomberg explained the PCE results:

Low unemployment, pandemic-era savings and wage growth are providing Americans the wherewithal to keep spending, allowing the economy to power ahead. Many economists have had to push out their recession calls, or in some cases, scrap them altogether as a result. The latest figures point to a strong start to economic growth in the third quarter… The saving rate decreased to 3.5%, the lowest since November, suggesting the recent pace of spending may not be sustained in coming months. Pandemic-era savings are dwindling, the labor market is cooling, delinquency rates are rising and the resumption of student loan payments this fall threatens to further strain consumers’ finances. Retail executives have suggested as much and remain cautious about the outlook.

If Bloomberg is right I should be able to keep using flat or falling into the Autumn.

Roughly 14,000 residents of west Maui have been seriously impacted by recent wildfires, ranging from total property-loss to sudden death. More than 100 deaths have been confirmed while it is widely assumed many more were killed in the conflagration. There are now 388 who have been validated as missing (updated estimates here). More than 10,000 have registered with FEMA for individual assistance (here). Maui’s total population is about 165,000.

As of yesterday, over 2900 people continue in emergency housing (hotels, Airbnb, and related, more and more and more). Many others are residing with family or friends. Rebuilding of housing stock will be complicated, delayed, extended, and expensive. At least 1500 residential properties have been destroyed, out of about 73,000 in Maui County. A significant portion of emergency housing has been secured using hotels and timeshares within one to two miles north of the Lahaina burn zone.

Unsafe Water Advisories continue for Lahaina, Upper Kula, and some other areas. Initial testing for dangerous chemicals has been undertaken (results). Further testing needs to be completed before public water system flows can support human hydration. Even bathing and cleaning should be avoided if possible (more). Both bulk and bottled potable water are being otherwise distributed.

Retail grocery outlets, including Foodland, Safeway, and Walmart are open and operating within twenty-miles of fire decimated Lahaina. SNAP benefits for hot food purchases are now authorized. A SNAP benefit to replace food lost in the disaster is available. The Lahaina Safeway survived intact and is ready to reopen when local officials give the okay. The Maui Food Bank has expanded its operations. Maui County and other emergency food/water distribution sites have been active. The Salvation Army, American Red Cross, and other voluntary organizations and individuals have been providing hot foods and other support. (More.)

Cell communications have mostly been restored (here and here). Other network connections have also been deployed (here and here). Here’s the FCC August 25 Communications Status Report.

Just under 1900 customers remain disconnected from the grid. In many cases a structure has not survived to be reconnected. More than 2200 structures have been seriously damaged or destroyed.

As previously posted here, the upstream and midstream supply chains for food, bottled water, and other essential freight have continued to operate and increased flows into Maui since the August 8 disaster (more and more and more).

In terms of Supply Chain Resilience, there is much more volume but much less velocity compared to August 7. Given the distance between Maui and its sources, for example, higher volume food flows are moving at about the same speed as before the disaster. But direction — more to the point, targeting — is much less precise since so many consumers are dispersed and some key nodes (such as the Lahaina Safeway) are not yet reopened.

This is not unusual.  This is typical.  It is common to mistake a velocity-problem for a volume-problem.  But not only do “solutions” for a volume problem not fix velocity, the added volume almost always increases congestion that further degrades velocity (here).  This is such a common mistake it has become close to a principle of Supply Chain Resilience diagnosis and prescription of post-disaster conditions.

Correction/Clarification: For my original post I misunderstood a report on emergency housing. The post above is corrected. The last group shelters (sometimes congregate shelters) have been closed. Displaced people are now housed in a range of facilities that will be paid-for using FEMA and state funding. This housing and financial support will persist for an extended period. Here is a helpful summary from MauiNow: “On Tuesday, Aug. 22, the last of the sheltered wildfire survivors staying at the South Maui Community Park Gymnasium in Kīhei, regained privacy as they were transitioned to longer-term housing. Since Aug. 16, the American Red Cross, Hawaiʻi Emergency Management Agency and the Federal Emergency Management Agency relocated nearly 3,000 survivors of the Maui wildfires from congregate shelters to contracted hotels. In total, more than a dozen hotels and timeshare properties are currently housing more than 4,400 shelter survivors and hotel employee survivors, while some 900 people are booked in Airbnb units. These accommodations will continue to provide meals, casework services, financial recovery resources and other disaster relief assistance.”

The Midcontinent Independent System Operator (MISO) is the regional reliability organization for electric utilities serving more than 40 million people in the central United States. Yesterday MISO released a “Maximum Generation Emergency Alert effective from 08/24/2023 12:00 EST until 08/24/2023 22:00 EST…” [See several updates below]

S&P Global reports, “Sweltering heat continues to push up power prices in the Central US, where the Southwest Power Pool set a peakload record of 56.18 GW Aug. 21 and the Midcontinent Independent System Operator was forecasting a record-smashing 130 GW of demand Aug. 24. In preparation for the soaring demand, MISO notified all local utilities in its footprint to prepare every available resource to serve the projected load, the grid operator said in a statement late Aug. 21. “We anticipate challenging operating conditions throughout the entire week, and we will need every available resource at some point,” Jessica Lucas, MISO executive director for system operations, said in the statement. (Below see SPG chart of MISO data.)

The emerging situation is of particular concern given the forward-looking assessment of MISO’s capacity in this year’s NERC Summer Reliability Assessment, “Demand forecasts and preliminary resource data indicate that MISO is at risk of operating reserve shortfalls during periods of high demand or low resource output. MISO’s resources are projected to be lower than in the summer of 2022 while net internal demand has also decreased. Firm transmission imports for this summer have significantly increased; this has resulted in a higher Anticipated Reserve Margin (ARM) of 23% (on an installed capacity basis) compared to 21% last summer…”

The prior paragraph is purposeful in highlighting ambivalence — both strengths and weaknesses. Bottom line: capacity is lower than last year and tomorrow’s demand is expected to spike. Our situation on Thursday night may mostly depend on some consumer restraint in the face of brutal temperatures and how the wind blows. The Summer Reliability Assessment for MISO also notes, “Performance of wind generators during periods of high electricity demand is a key factor in determining whether system operators need to employ operating mitigations, such as maximum-generation declarations and energy emergencies. MISO has over 30,300 MW of installed wind capacity; however, the historically-based on-peak capacity contribution is 5,488 MW.”

The MISO demand/supply dashboard can be accessed here.

+++

August 24 Update: The following alert was distributed 8:16 Central Time this morning: “MISO is escalating the Maximum Generation Alert to a Maximum Generation Emergency Event Step 2a effective from 12:00  until 22:00 EST today for the MISO Balancing Authority Area. The reason for the Event is because of Forced Generation Outages, Above Normal Temps, Higher than Forecasted Load.” According to Reuters, “MISO projected it would have 125,907 MW of supplies available with 120,656 MW from internal resources and 5,251 MW of imports. That would not be enough to meet the grid’s forecast peak of 127,692 MW, which would top the system’s all-time high of 127,100 MW in July 2011.” Reuters and others are also reporting that the Texas grid is facing similar shortfalls today (more).

Later: At 2:00 Central Time on Thursday some personal observations: MISO’s solar and fossil-fuel generators are achieving forecasts. MISO’s wind generation has been less than forecast most of today. For example, the day-ahead forecast for 1:00 Central anticipated 4294MW being generated, instead 3134.33MW was being generated. At 2:00 the deficit had grown to just over 1500MW. Imports seem to be arriving as anticipated. This is, however, a wide-area challenge encompassing at least the Southwest Power Pool, ERCOT, and MISO. Will exports from other regions be sufficient to fill gaps across such a wide area? Behavior of the national grid can be monitored on the EIA website.

August 25 Early Morning Update: The mid-continent’s grid did not fail on Thursday. Both MISO and ERCOT avoided rolling blackouts and other extreme measures, but barely. Good news does not attract the attention that would have been given the opposite. I will be looking for more detailed assessments from domain specialists. Bloomberg is crediting a pop-up thunderstorm for cooling demand in Houston. Right now, it seems to me that in combination with other measures, it was possible to import more power from other regions when it was needed most. The chart immediately below is most of yesterday’s import pattern for MISO. At about 4:45 PM local time (16:45 on chart) imports peaked (or on this chart bottomed-out) at slightly more than 8745MW, even more than scheduled, much more than expected on Thursday morning. I won’t have access to data for another 24 hours or so, but I bet PJM was exporting lots of megawatts to MISO. In the case of ERCOT (Texas), they were importing too, but I am as intrigued with the release of “stored power”(see second chart below). Just before 8PM local this was contributing 1132MW… just as solar power was fading and evening demand was peaking. Related to this report? I’m not sure. Almost certainly related to this report.

Below is the August 24 Fuel Mix for ERCOT (Electric Reliability Council of Texas)

August 25 Afternoon Update: Bloomberg reports, “A punishing heat wave will test power supplies in Texas for a second day as triple-digit temperatures continue to drive up electricity consumption. The state’s grid operator expects to have almost 80 gigawatts of capacity available at about 8 p.m. local time, after solar farms start shutting off, based on hourly forecasts. Demand will be within about 1 gigawatt of that total. Five-minute forecasts show demand exceeding supply around that time…”

August 26 Update: It was once again very close, but the Texas grid made it through another sundown. Here’s more from the Dallas NBC affiliate. Below are the Friday, August 25 supply slopes. Excessive heat is forecast to continue through this weekend. MISO has also declared a Hot Weather Alert through Sunday.

August 27 Update: The Texas grid continues to operate along the cusp. Reuters is reporting renewed conservation requests by ERCOT. The warnings and earnest requests are prudent. Repeated skin-of-the-teeth avoidance of blackouts does tend to reduce attention and related conservation behavior…

September 2 Update: S&P Global has published an important, helpful review of ERCOT performance over the month of August. The report includes the following:

ERCOT President and CEO Pablo Vegas said Texas power consumers’ response to requests to conserve energy enabled the grid to avoid declaring Energy Emergency Alerts so far this summer, despite setting 10 peakload records, most recently 85.4 GW on Aug. 10. In comparison with summer 2022’s all-time peak of 80.1 GW, the increase has been more than 5 GW, Vegas said, about 7% growth year on year, compared with nationwide annual load growth averaging about 1% since the early 2000s. Texas’ energy demand reflects the growing Texas population and economy, which bring with them “demands on infrastructure in highways, schools and energy,” Vegas said. “I don’t think anybody expects the growth to slow down meaningfully…”

An August 31 report in the Texas Tribune mostly agrees with Vargas and raises the threat of conservation fatigue. “Repetitive power conservation appeals by the state’s grid operator face a challenge: Texans becoming less responsive to calls. Past conservation requests have helped reduce 100,000 homes worth of power demand on the grid.”

Freight is where, when, and how push meets pull. Demand prompts supply, freight fulfills demand. Air cargo, maritime (riverine and ocean-going), train, truckload, LTL, refrigerated, van, flatbed, package delivery, and even more reflect diverse modes, intermodal possibilities, and diverse, responsive spot and contract markets. Freight is most of how volumes achieve velocity.

The many parts of the freight sector are highly interdependent, but most deliveries finally depend on some sort of truck and trucker on the road able to refuel (here and here).

The chart below suggests the steep peaks, deep valleys, and recurring volatility of the US trucking market. The red line is basically volume (compared to a 1990 baseline). The blue line is the Producer Price Index (PPI) for the trucking sector, measuring price changes freight carriers command or suffer. When demand for stuff is strong and stresses capacity (e.g., first half of 2022), prices will trend higher. When freight demand falls, especially if capacity hangs on (e.g., since at least last summer), prices will be softer… in this case, compared to a 1992 base-line (more and more).

There are plenty of other freight indicators: FreightWaves SONAR, DAT Trendlines, Bank of America Truckload Demand Indicator, U.S. Bank Freight Payments Index, ATA Truck Tonnage Index — each offering different angles more or less current. Diesel prices are absolutely worth tracking. Most of these smoke signals now agree — some more, some less enthusiastically — with Brooke Sutherland at Bloomberg “The free-falling freight market may finally be finding a bottom. Shippers shouldn’t count on a return to the go-go days of Covid, but resilient US consumers offer a way out of the morass…”

Volumes are more or less fine, not great but no where near the deepest valleys of 2020 or even 2017. Price-drops may have stopped (at least slowed) while still above pre-pandemic levels, even after adjusting for inflation (see second chart below). The recent loss or downsizing of some trucking firms, such as Yellow, can be good news for those still hauling.

Freight movement is analogous to human blood pressure, not too high and not too low is best for the whole supply chain. I’m sure truckers hope our current pace and related pressure builds a bit, there is existing capacity to do more. But current flow volumes and velocity are inside healthy ranges.

One Month Percent Change: PPI Transportation Services-truck transportation of freight, not seasonally adjusted

Bureau of Labor Statistics, Data extracted on: August 17, 2023 (5:59:47 AM)

Upstream and midstream flows serving Maui are operating above preexisting capacity. Downstream flows — and especially demand velocities — have been very seriously disrupted. What is needed, how much is needed, and precisely where it is needed are not yet entirely clear to downstream operators.

Upstream flows from Port of Oakland and Port of Long Beach are consistent with prior schedules. Matson and Pasha are the principal Jones Act carriers delivering groceries, medical goods, and other essentials into Honolulu. Air cargo flights into Hawaii and, specifically, into Kahului Airport on Maui have been expanded (here and here and here).

Midstream flows from Honolulu to Maui have also been consistent with or better than prior volumes and velocity. Maui is mostly supplied out of Honolulu (less than 100 miles distant, one way ferry transport usually takes 2.5 hours). Here is how Matson explains their current flow:

Matson operates a hub-and-spoke system in Hawaii, sailing mainline containerships between the U.S. West Coast and Honolulu three times a week, with connecting service to the Neighbor Islands on owned barges. It will maintain its normal schedule of barge arrivals at Kahului on Tuesdays and Fridays. The company has scheduled one extra Maui sailing, on Sunday, August 20, and is considering additional sailings to Maui should they be needed.

Young Brothers is an especially important midstream — inter-island — carrier. As early as August 9, Young Brothers explained:

The Port of Kahului is fully operational and will be accepting and delivering cargo as normal. To support relief efforts in the face of the disastrous wildfires that have impacted the communities of Maui, we will be prioritizing the loading and discharge of medical supplies and emergency and utility vehicles to the port of Kahului over the next few sailings. Young Brothers will strive to ensure that bookings for gated cargo sail on the currently scheduled sailing date(s). However, due to shifts in capacity for certain sailings, Young Brothers may re-book some cargo to the next available sailing.  As a reminder, less than container load cargo is moved on a space available basis and priority will be given to cargo that support relief efforts.

Downstream flows, especially in western Maui (see maps below and here), suddenly fractured (in some places melted or were reduced to ash), consumer ability to signal demand was very constrained even as immediate demand spiked. Demand nodes — such as grocery stores and restaurants — have been destroyed. The ability to purchase supply has been reduced by loss of these nodes, unreliable cell networks, and loss of related digital transaction networks. Most housing in Lahaina (population: @ 13,000) is estimated to be a total loss, including the loss of residential food stocks (see second map below). Most supply channels have survived, but localized scope and scale were sparse even before this disaster and perimeter controls remain in place. The local public water system is mostly intact, but in the hardest hit locations water is considered unsafe to drink. Grid power remains unavailable to roughly 5000 customers in Maui County. Cell coverage is beginning to recover.

The flow networks supplying Maui’s demand were not seriously impacted by the August 8 and subsequent wildfires. These flow networks have flexed in response to devastating local losses. But western Maui, in particular, is on the far outer edge of high volume, high velocity demand and supply networks. In the case of no-notice or fast-onset major disasters, local demand will often be separated from outside flow for 72 hours or longer. Local demand is usually displaced and discombobulated. Local supply nodes are often disrupted or destroyed. Reconnecting network supply with local demand is delayed by lack of communication, loss of essential infrastructure, and many other factors. Utter destruction of the built environment, as in Lahaina, will extend this isolation period… especially where preexisting flows were already peripheral.

Initial response, inside this 72 hour window, is most likely to spontaneously emerge from immediately proximate and less-systematic behavior where survivors directly help survivors (more and more). The more physically isolated the disaster locale, the more time-extended this period of initial response. The more proximate the impact zone to a comparatively unscathed concentration of supply, the more quickly network capacity can be recovered and deployed. None of this is surprising. All of this has been empirically demonstrated and confirmed for over seventy years (one classic example here).

The recurrent reality of this delay is not comforting to survivors. Many disaster response professionals (both private and public) work hard to reduce this delay. But physical limitations of distance, time, capacity, and interdependencies can be very stubborn. More comforting is the “cornucopia” of supply that typically emerges in the extended aftermath of the isolation period… at least where and when effectual demand persists and preexisting network capacity has also survived.

[On August 25 a new post updated the status of recovery for Maui.]